diff --git a/build/cpu.o b/build/cpu.o
index 6228ac2..6429509 100644
Binary files a/build/cpu.o and b/build/cpu.o differ
diff --git a/build/cuda.cu.o b/build/cuda.cu.o
index 376fab2..60b85e1 100644
Binary files a/build/cuda.cu.o and b/build/cuda.cu.o differ
diff --git a/build/tensor.o b/build/tensor.o
index d330a48..10402c8 100644
Binary files a/build/tensor.o and b/build/tensor.o differ
diff --git a/examples/train.ipynb b/examples/train.ipynb
index 3a37c54..b29a29d 100644
--- a/examples/train.ipynb
+++ b/examples/train.ipynb
@@ -33,16 +33,19963 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Epoch [1/10], Loss: 1.7035\n",
-      "Epoch [2/10], Loss: 0.7193\n",
-      "Epoch [3/10], Loss: 0.3068\n",
-      "Epoch [4/10], Loss: 0.1742\n",
-      "Epoch [5/10], Loss: 0.1342\n",
-      "Epoch [6/10], Loss: 0.1232\n",
-      "Epoch [7/10], Loss: 0.1220\n",
-      "Epoch [8/10], Loss: 0.1241\n",
-      "Epoch [9/10], Loss: 0.1270\n",
-      "Epoch [10/10], Loss: 0.1297\n"
+      "Download train-images-idx3-ubyte.gz from http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to .data/mnist/train-images-idx3-ubyte.gz\n",
+      "Downloading... 100% | [==================================================] | Done !\n",
+      ".data/mnist/mnist-data-py\n",
+      "Extracting... | NOTE: gzip files are not extracted, and moved to .data/mnist/mnist-data-py | Done !\n",
+      "Download train-labels-idx1-ubyte.gz from http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to .data/mnist/train-labels-idx1-ubyte.gz\n",
+      "Downloading... 100% | [==================================================] | Done !\n",
+      ".data/mnist/mnist-data-py\n",
+      "Extracting... | NOTE: gzip files are not extracted, and moved to .data/mnist/mnist-data-py | Done !\n",
+      "Download t10k-labels-idx1-ubyte.gz from http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to .data/mnist/t10k-labels-idx1-ubyte.gz\n",
+      "Downloading... 100% | [==================================================] | Done !\n",
+      ".data/mnist/mnist-data-py\n",
+      "Extracting... | NOTE: gzip files are not extracted, and moved to .data/mnist/mnist-data-py | Done !\n",
+      "Download t10k-images-idx3-ubyte.gz from http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to .data/mnist/t10k-images-idx3-ubyte.gz\n",
+      "Downloading... 100% | [==================================================] | Done !\n",
+      ".data/mnist/mnist-data-py\n",
+      "Extracting... | NOTE: gzip files are not extracted, and moved to .data/mnist/mnist-data-py | Done !\n"
+     ]
+    },
+    {
+     "ename": "BadGzipFile",
+     "evalue": "Not a gzipped file (b'<!')",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[0;31mBadGzipFile\u001b[0m                               Traceback (most recent call last)",
+      "Cell \u001b[0;32mIn[2], line 6\u001b[0m\n\u001b[1;32m      3\u001b[0m reshape \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;01mlambda\u001b[39;00m x: x\u001b[38;5;241m.\u001b[39mreshape(\u001b[38;5;241m-\u001b[39m\u001b[38;5;241m1\u001b[39m, \u001b[38;5;241m784\u001b[39m)\n\u001b[1;32m      5\u001b[0m \u001b[38;5;66;03m# Download the training and testing data\u001b[39;00m\n\u001b[0;32m----> 6\u001b[0m train_data, test_data \u001b[38;5;241m=\u001b[39m \u001b[43mnorch\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mdatasets\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mMNIST\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43msplits\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtransform\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43m \u001b[49m\u001b[43mreshape\u001b[49m\u001b[43m)\u001b[49m\n",
+      "File \u001b[0;32m~/Documentos/recreate_pytorch/PyNorch/norch/datasets/mnist.py:75\u001b[0m, in \u001b[0;36mMNIST.splits\u001b[0;34m(cls, root, train_data, train_label, test_data, test_label, **kwargs)\u001b[0m\n\u001b[1;32m     73\u001b[0m test_data \u001b[38;5;241m=\u001b[39m os\u001b[38;5;241m.\u001b[39mpath\u001b[38;5;241m.\u001b[39mjoin(path, test_data)\n\u001b[1;32m     74\u001b[0m test_label \u001b[38;5;241m=\u001b[39m os\u001b[38;5;241m.\u001b[39mpath\u001b[38;5;241m.\u001b[39mjoin(path, test_label)\n\u001b[0;32m---> 75\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mMNIST\u001b[49m\u001b[43m(\u001b[49m\u001b[43mtrain_data\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mtrain_label\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m, MNIST(test_data, test_label, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mkwargs)\n",
+      "File \u001b[0;32m~/Documentos/recreate_pytorch/PyNorch/norch/datasets/mnist.py:39\u001b[0m, in \u001b[0;36mMNIST.__init__\u001b[0;34m(self, path_data, path_label, transform)\u001b[0m\n\u001b[1;32m     38\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m__init__\u001b[39m(\u001b[38;5;28mself\u001b[39m, path_data, path_label, transform\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mNone\u001b[39;00m):\n\u001b[0;32m---> 39\u001b[0m     data \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_load_mnist\u001b[49m\u001b[43m(\u001b[49m\u001b[43mpath_data\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mheader_size\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;241;43m16\u001b[39;49m\u001b[43m)\u001b[49m\u001b[38;5;241m.\u001b[39mreshape((\u001b[38;5;241m-\u001b[39m\u001b[38;5;241m1\u001b[39m, \u001b[38;5;241m28\u001b[39m, \u001b[38;5;241m28\u001b[39m))\n\u001b[1;32m     40\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m transform \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[1;32m     41\u001b[0m         data \u001b[38;5;241m=\u001b[39m transform(data)\n",
+      "File \u001b[0;32m~/Documentos/recreate_pytorch/PyNorch/norch/datasets/mnist.py:47\u001b[0m, in \u001b[0;36mMNIST._load_mnist\u001b[0;34m(self, path, header_size)\u001b[0m\n\u001b[1;32m     45\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m_load_mnist\u001b[39m(\u001b[38;5;28mself\u001b[39m, path, header_size):\n\u001b[1;32m     46\u001b[0m     \u001b[38;5;28;01mwith\u001b[39;00m gzip\u001b[38;5;241m.\u001b[39mopen(path, \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mrb\u001b[39m\u001b[38;5;124m'\u001b[39m) \u001b[38;5;28;01mas\u001b[39;00m f:\n\u001b[0;32m---> 47\u001b[0m         \u001b[43mf\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mseek\u001b[49m\u001b[43m(\u001b[49m\u001b[43mheader_size\u001b[49m\u001b[43m)\u001b[49m  \u001b[38;5;66;03m# Skip the header\u001b[39;00m\n\u001b[1;32m     48\u001b[0m         data \u001b[38;5;241m=\u001b[39m f\u001b[38;5;241m.\u001b[39mread()\n\u001b[1;32m     50\u001b[0m     data_list \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mlist\u001b[39m(data)\n",
+      "File \u001b[0;32m/usr/lib/python3.8/gzip.py:384\u001b[0m, in \u001b[0;36mGzipFile.seek\u001b[0;34m(self, offset, whence)\u001b[0m\n\u001b[1;32m    382\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mmode \u001b[38;5;241m==\u001b[39m READ:\n\u001b[1;32m    383\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_check_not_closed()\n\u001b[0;32m--> 384\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_buffer\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mseek\u001b[49m\u001b[43m(\u001b[49m\u001b[43moffset\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mwhence\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    386\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39moffset\n",
+      "File \u001b[0;32m/usr/lib/python3.8/_compression.py:143\u001b[0m, in \u001b[0;36mDecompressReader.seek\u001b[0;34m(self, offset, whence)\u001b[0m\n\u001b[1;32m    141\u001b[0m \u001b[38;5;66;03m# Read and discard data until we reach the desired position.\u001b[39;00m\n\u001b[1;32m    142\u001b[0m \u001b[38;5;28;01mwhile\u001b[39;00m offset \u001b[38;5;241m>\u001b[39m \u001b[38;5;241m0\u001b[39m:\n\u001b[0;32m--> 143\u001b[0m     data \u001b[38;5;241m=\u001b[39m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mread\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mmin\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43mio\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mDEFAULT_BUFFER_SIZE\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43moffset\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    144\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m data:\n\u001b[1;32m    145\u001b[0m         \u001b[38;5;28;01mbreak\u001b[39;00m\n",
+      "File \u001b[0;32m/usr/lib/python3.8/gzip.py:479\u001b[0m, in \u001b[0;36m_GzipReader.read\u001b[0;34m(self, size)\u001b[0m\n\u001b[1;32m    475\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_new_member:\n\u001b[1;32m    476\u001b[0m     \u001b[38;5;66;03m# If the _new_member flag is set, we have to\u001b[39;00m\n\u001b[1;32m    477\u001b[0m     \u001b[38;5;66;03m# jump to the next member, if there is one.\u001b[39;00m\n\u001b[1;32m    478\u001b[0m     \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_init_read()\n\u001b[0;32m--> 479\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_read_gzip_header\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m:\n\u001b[1;32m    480\u001b[0m         \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_size \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_pos\n\u001b[1;32m    481\u001b[0m         \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;124mb\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m\"\u001b[39m\n",
+      "File \u001b[0;32m/usr/lib/python3.8/gzip.py:427\u001b[0m, in \u001b[0;36m_GzipReader._read_gzip_header\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    424\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;01mFalse\u001b[39;00m\n\u001b[1;32m    426\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m magic \u001b[38;5;241m!=\u001b[39m \u001b[38;5;124mb\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;130;01m\\037\u001b[39;00m\u001b[38;5;130;01m\\213\u001b[39;00m\u001b[38;5;124m'\u001b[39m:\n\u001b[0;32m--> 427\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m BadGzipFile(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mNot a gzipped file (\u001b[39m\u001b[38;5;132;01m%r\u001b[39;00m\u001b[38;5;124m)\u001b[39m\u001b[38;5;124m'\u001b[39m \u001b[38;5;241m%\u001b[39m magic)\n\u001b[1;32m    429\u001b[0m (method, flag,\n\u001b[1;32m    430\u001b[0m  \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_last_mtime) \u001b[38;5;241m=\u001b[39m struct\u001b[38;5;241m.\u001b[39munpack(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m<BBIxx\u001b[39m\u001b[38;5;124m\"\u001b[39m, \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_read_exact(\u001b[38;5;241m8\u001b[39m))\n\u001b[1;32m    431\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m method \u001b[38;5;241m!=\u001b[39m \u001b[38;5;241m8\u001b[39m:\n",
+      "\u001b[0;31mBadGzipFile\u001b[0m: Not a gzipped file (b'<!')"
+     ]
+    }
+   ],
+   "source": [
+    "import norch\n",
+    "\n",
+    "to_tensor = lambda x: norch.Tensor(x)\n",
+    "reshape = lambda x: x.reshape([-1, 784])\n",
+    "transform = lambda x: reshape(to_tensor(x))\n",
+    "\n",
+    "# Download the training and testing data\n",
+    "train_data, test_data = norch.datasets.MNIST.splits(transform = reshape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "loss_antes tensor([3.3399136066436768,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6715301871299744,],\n",
+      "[-0.13446585834026337,],\n",
+      "[0.5245601534843445,],\n",
+      "[-0.9957879185676575,],\n",
+      "[-0.10922560840845108,],\n",
+      "[0.4430800676345825,],\n",
+      "[-0.5424755811691284,],\n",
+      "[0.8905413746833801,],\n",
+      "[0.8028548955917358,],\n",
+      "[-0.9388200044631958,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.7763985395431519,],\n",
+      "[-0.07534199953079224,],\n",
+      "[-0.7497910857200623,],\n",
+      "[0.17107321321964264,],\n",
+      "[-0.05041635036468506,],\n",
+      "[-0.020194709300994873,],\n",
+      "[-0.1415593922138214,],\n",
+      "[-0.1368662416934967,],\n",
+      "[-0.02720409631729126,],\n",
+      "[-0.002188590820878744,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5338311195373535,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.6550822257995605,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6707537770271301,],\n",
+      "[-0.13439051806926727,],\n",
+      "[0.5253099203109741,],\n",
+      "[-0.9959589838981628,],\n",
+      "[-0.1091751903295517,],\n",
+      "[0.4431002736091614,],\n",
+      "[-0.5423340201377869,],\n",
+      "[0.8906782269477844,],\n",
+      "[0.8028820753097534,],\n",
+      "[-0.9388177990913391,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.530176043510437,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.5491833686828613,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6707537770271301,],\n",
+      "[-0.13439051806926727,],\n",
+      "[0.5253099203109741,],\n",
+      "[-0.9959589838981628,],\n",
+      "[-0.1091751903295517,],\n",
+      "[0.4431002736091614,],\n",
+      "[-0.5423340201377869,],\n",
+      "[0.8906782269477844,],\n",
+      "[0.8028820753097534,],\n",
+      "[-0.9388177990913391,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.8605958223342896,],\n",
+      "[-0.07921524345874786,],\n",
+      "[-0.7264853715896606,],\n",
+      "[0.15928134322166443,],\n",
+      "[-0.034484878182411194,],\n",
+      "[-0.014619462192058563,],\n",
+      "[-0.09272399544715881,],\n",
+      "[-0.1671433448791504,],\n",
+      "[-0.013369085267186165,],\n",
+      "[-0.0027314182370901108,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.530176043510437,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.7678496837615967,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.669893205165863,],\n",
+      "[-0.13431130349636078,],\n",
+      "[0.5260363817214966,],\n",
+      "[-0.9961182475090027,],\n",
+      "[-0.10914070904254913,],\n",
+      "[0.44311490654945374,],\n",
+      "[-0.5422412753105164,],\n",
+      "[0.8908453583717346,],\n",
+      "[0.8028954267501831,],\n",
+      "[-0.9388150572776794,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5264081954956055,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([4.617180824279785,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.669893205165863,],\n",
+      "[-0.13431130349636078,],\n",
+      "[0.5260363817214966,],\n",
+      "[-0.9961182475090027,],\n",
+      "[-0.10914070904254913,],\n",
+      "[0.44311490654945374,],\n",
+      "[-0.5422412753105164,],\n",
+      "[0.8908453583717346,],\n",
+      "[0.8028954267501831,],\n",
+      "[-0.9388150572776794,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.012890100479126,],\n",
+      "[-0.08897354453802109,],\n",
+      "[-0.7638409733772278,],\n",
+      "[0.16191589832305908,],\n",
+      "[0.03984580934047699,],\n",
+      "[0.006034770980477333,],\n",
+      "[0.04748964309692383,],\n",
+      "[-0.16453632712364197,],\n",
+      "[0.00813131034374237,],\n",
+      "[-0.0028574985917657614,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5264081954956055,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-4.297521591186523,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6688803434371948,],\n",
+      "[-0.13422232866287231,],\n",
+      "[0.5268002152442932,],\n",
+      "[-0.9962801337242126,],\n",
+      "[-0.10918055474758148,],\n",
+      "[0.4431088864803314,],\n",
+      "[-0.5422887802124023,],\n",
+      "[0.8910098671913147,],\n",
+      "[0.8028873205184937,],\n",
+      "[-0.9388121962547302,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5221107006072998,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([5.775960922241211,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6688803434371948,],\n",
+      "[-0.13422232866287231,],\n",
+      "[0.5268002152442932,],\n",
+      "[-0.9962801337242126,],\n",
+      "[-0.10918055474758148,],\n",
+      "[0.4431088864803314,],\n",
+      "[-0.5422887802124023,],\n",
+      "[0.8910098671913147,],\n",
+      "[0.8028873205184937,],\n",
+      "[-0.9388121962547302,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.119827151298523,],\n",
+      "[-0.09500832855701447,],\n",
+      "[-0.7736548185348511,],\n",
+      "[0.1594274789094925,],\n",
+      "[0.11072523146867752,],\n",
+      "[0.025451630353927612,],\n",
+      "[0.17995238304138184,],\n",
+      "[-0.17571359872817993,],\n",
+      "[0.022571373730897903,],\n",
+      "[-0.0030716522596776485,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5221107006072998,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-4.806645393371582,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6677604913711548,],\n",
+      "[-0.13412731885910034,],\n",
+      "[0.5275738835334778,],\n",
+      "[-0.9964395761489868,],\n",
+      "[-0.10929127782583237,],\n",
+      "[0.44308343529701233,],\n",
+      "[-0.5424687266349792,],\n",
+      "[0.8911855816841125,],\n",
+      "[0.8028647303581238,],\n",
+      "[-0.9388090968132019,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5173040628433228,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([5.9341840744018555,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6677604913711548,],\n",
+      "[-0.13412731885910034,],\n",
+      "[0.5275738835334778,],\n",
+      "[-0.9964395761489868,],\n",
+      "[-0.10929127782583237,],\n",
+      "[0.44308343529701233,],\n",
+      "[-0.5424687266349792,],\n",
+      "[0.8911855816841125,],\n",
+      "[0.8028647303581238,],\n",
+      "[-0.9388090968132019,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.0746594667434692,],\n",
+      "[-0.08936231583356857,],\n",
+      "[-0.6985778212547302,],\n",
+      "[0.14072144031524658,],\n",
+      "[0.11916185170412064,],\n",
+      "[0.02836163155734539,],\n",
+      "[0.20278218388557434,],\n",
+      "[-0.22442466020584106,],\n",
+      "[0.016736404970288277,],\n",
+      "[-0.0034379507414996624,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5173040628433228,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-4.872036457061768,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6666858196258545,],\n",
+      "[-0.13403795659542084,],\n",
+      "[0.5282724499702454,],\n",
+      "[-0.9965803027153015,],\n",
+      "[-0.10941044241189957,],\n",
+      "[0.44305506348609924,],\n",
+      "[-0.5426715016365051,],\n",
+      "[0.8914099931716919,],\n",
+      "[0.8028479814529419,],\n",
+      "[-0.9388056397438049,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5124320387840271,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([4.719270706176758,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6666858196258545,],\n",
+      "[-0.13403795659542084,],\n",
+      "[0.5282724499702454,],\n",
+      "[-0.9965803027153015,],\n",
+      "[-0.10941044241189957,],\n",
+      "[0.44305506348609924,],\n",
+      "[-0.5426715016365051,],\n",
+      "[0.8914099931716919,],\n",
+      "[0.8028479814529419,],\n",
+      "[-0.9388056397438049,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.8705560564994812,],\n",
+      "[-0.07186680287122726,],\n",
+      "[-0.546619713306427,],\n",
+      "[0.10827986896038055,],\n",
+      "[0.04532695934176445,],\n",
+      "[0.010129470378160477,],\n",
+      "[0.07962340116500854,],\n",
+      "[-0.30150407552719116,],\n",
+      "[-0.008896633982658386,],\n",
+      "[-0.0038197529502213,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5124320387840271,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-4.344776630401611,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6658152341842651,],\n",
+      "[-0.1339660882949829,],\n",
+      "[0.5288190841674805,],\n",
+      "[-0.9966886043548584,],\n",
+      "[-0.10945577174425125,],\n",
+      "[0.4430449306964874,],\n",
+      "[-0.542751133441925,],\n",
+      "[0.8917114734649658,],\n",
+      "[0.8028568625450134,],\n",
+      "[-0.9388018250465393,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5080872774124146,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.9647116661071777,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6658152341842651,],\n",
+      "[-0.1339660882949829,],\n",
+      "[0.5288190841674805,],\n",
+      "[-0.9966886043548584,],\n",
+      "[-0.10945577174425125,],\n",
+      "[0.4430449306964874,],\n",
+      "[-0.542751133441925,],\n",
+      "[0.8917114734649658,],\n",
+      "[0.8028568625450134,],\n",
+      "[-0.9388018250465393,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.6034453511238098,],\n",
+      "[-0.04986261576414108,],\n",
+      "[-0.3749369978904724,],\n",
+      "[0.07348798215389252,],\n",
+      "[-0.0794970691204071,],\n",
+      "[-0.02075963281095028,],\n",
+      "[-0.13515466451644897,],\n",
+      "[-0.3795510530471802,],\n",
+      "[-0.0413801446557045,],\n",
+      "[-0.004105218220502138,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5080872774124146,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.4436697959899902,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6652117967605591,],\n",
+      "[-0.1339162290096283,],\n",
+      "[0.5291939973831177,],\n",
+      "[-0.9967620968818665,],\n",
+      "[-0.10937627404928207,],\n",
+      "[0.44306570291519165,],\n",
+      "[-0.5426159501075745,],\n",
+      "[0.8920910358428955,],\n",
+      "[0.8028982281684875,],\n",
+      "[-0.9387977123260498,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5046436190605164,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.7221570014953613,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6652117967605591,],\n",
+      "[-0.1339162290096283,],\n",
+      "[0.5291939973831177,],\n",
+      "[-0.9967620968818665,],\n",
+      "[-0.10937627404928207,],\n",
+      "[0.44306570291519165,],\n",
+      "[-0.5426159501075745,],\n",
+      "[0.8920910358428955,],\n",
+      "[0.8028982281684875,],\n",
+      "[-0.9387977123260498,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.38911694288253784,],\n",
+      "[-0.03231014311313629,],\n",
+      "[-0.24431681632995605,],\n",
+      "[0.047654666006565094,],\n",
+      "[-0.19218146800994873,],\n",
+      "[-0.04830428957939148,],\n",
+      "[-0.33135926723480225,],\n",
+      "[-0.4348355531692505,],\n",
+      "[-0.0667501762509346,],\n",
+      "[-0.004262998234480619,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5046436190605164,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.624621868133545,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6648226976394653,],\n",
+      "[-0.13388392329216003,],\n",
+      "[0.529438316822052,],\n",
+      "[-0.9968097805976868,],\n",
+      "[-0.10918409377336502,],\n",
+      "[0.4431140124797821,],\n",
+      "[-0.5422846078872681,],\n",
+      "[0.892525851726532,],\n",
+      "[0.802964985370636,],\n",
+      "[-0.938793420791626,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.5020189881324768,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.3282623291015625,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6648226976394653,],\n",
+      "[-0.13388392329216003,],\n",
+      "[0.529438316822052,],\n",
+      "[-0.9968097805976868,],\n",
+      "[-0.10918409377336502,],\n",
+      "[0.4431140124797821,],\n",
+      "[-0.5422846078872681,],\n",
+      "[0.892525851726532,],\n",
+      "[0.802964985370636,],\n",
+      "[-0.938793420791626,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.28115907311439514,],\n",
+      "[-0.0235289353877306,],\n",
+      "[-0.1815749853849411,],\n",
+      "[0.03541348874568939,],\n",
+      "[-0.2358240932226181,],\n",
+      "[-0.05850483477115631,],\n",
+      "[-0.4132963716983795,],\n",
+      "[-0.4600899815559387,],\n",
+      "[-0.0793987289071083,],\n",
+      "[-0.004317693877965212,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.5020189881324768,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.305006504058838,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6645415425300598,],\n",
+      "[-0.13386039435863495,],\n",
+      "[0.5296198725700378,],\n",
+      "[-0.9968451857566833,],\n",
+      "[-0.10894826799631119,],\n",
+      "[0.44317251443862915,],\n",
+      "[-0.5418713092803955,],\n",
+      "[0.8929859399795532,],\n",
+      "[0.8030443787574768,],\n",
+      "[-0.9387891292572021,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4997139871120453,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.703629970550537,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6645415425300598,],\n",
+      "[-0.13386039435863495,],\n",
+      "[0.5296198725700378,],\n",
+      "[-0.9968451857566833,],\n",
+      "[-0.10894826799631119,],\n",
+      "[0.44317251443862915,],\n",
+      "[-0.5418713092803955,],\n",
+      "[0.8929859399795532,],\n",
+      "[0.8030443787574768,],\n",
+      "[-0.9387891292572021,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.25594568252563477,],\n",
+      "[-0.021709876134991646,],\n",
+      "[-0.17255055904388428,],\n",
+      "[0.0337710864841938,],\n",
+      "[-0.1938377022743225,],\n",
+      "[-0.0474424846470356,],\n",
+      "[-0.35894909501075745,],\n",
+      "[-0.4635763168334961,],\n",
+      "[-0.0826754942536354,],\n",
+      "[-0.004316887818276882,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4997139871120453,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.610466957092285,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6642856001853943,],\n",
+      "[-0.1338386833667755,],\n",
+      "[0.5297924280166626,],\n",
+      "[-0.996878981590271,],\n",
+      "[-0.10875443369150162,],\n",
+      "[0.44321995973587036,],\n",
+      "[-0.5415123701095581,],\n",
+      "[0.8934495449066162,],\n",
+      "[0.8031270503997803,],\n",
+      "[-0.9387848377227783,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4971035122871399,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.719329595565796,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6642856001853943,],\n",
+      "[-0.1338386833667755,],\n",
+      "[0.5297924280166626,],\n",
+      "[-0.996878981590271,],\n",
+      "[-0.10875443369150162,],\n",
+      "[0.44321995973587036,],\n",
+      "[-0.5415123701095581,],\n",
+      "[0.8934495449066162,],\n",
+      "[0.8031270503997803,],\n",
+      "[-0.9387848377227783,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.2549159824848175,],\n",
+      "[-0.022134995087981224,],\n",
+      "[-0.18165110051631927,],\n",
+      "[0.03576912358403206,],\n",
+      "[-0.1000286340713501,],\n",
+      "[-0.023694530129432678,],\n",
+      "[-0.23029501736164093,],\n",
+      "[-0.45979487895965576,],\n",
+      "[-0.08331718295812607,],\n",
+      "[-0.004302910529077053,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4971035122871399,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.298074722290039,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6640306711196899,],\n",
+      "[-0.13381655514240265,],\n",
+      "[0.529974102973938,],\n",
+      "[-0.9969147443771362,],\n",
+      "[-0.10865440219640732,],\n",
+      "[0.4432436525821686,],\n",
+      "[-0.541282057762146,],\n",
+      "[0.8939093351364136,],\n",
+      "[0.8032103776931763,],\n",
+      "[-0.9387805461883545,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4938054382801056,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.8078713417053223,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6640306711196899,],\n",
+      "[-0.13381655514240265,],\n",
+      "[0.529974102973938,],\n",
+      "[-0.9969147443771362,],\n",
+      "[-0.10865440219640732,],\n",
+      "[0.4432436525821686,],\n",
+      "[-0.541282057762146,],\n",
+      "[0.8939093351364136,],\n",
+      "[0.8032103776931763,],\n",
+      "[-0.9387805461883545,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.2340373396873474,],\n",
+      "[-0.021054279059171677,],\n",
+      "[-0.17807479202747345,],\n",
+      "[0.03534828498959541,],\n",
+      "[-0.01827407069504261,],\n",
+      "[-0.0032431890722364187,],\n",
+      "[-0.12982770800590515,],\n",
+      "[-0.4598676264286041,],\n",
+      "[-0.08593430370092392,],\n",
+      "[-0.004297984763979912,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4938054382801056,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.9027509689331055,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6637966632843018,],\n",
+      "[-0.13379549980163574,],\n",
+      "[0.5301522016525269,],\n",
+      "[-0.996950089931488,],\n",
+      "[-0.10863612592220306,],\n",
+      "[0.44324690103530884,],\n",
+      "[-0.5411522388458252,],\n",
+      "[0.8943691849708557,],\n",
+      "[0.8032963275909424,],\n",
+      "[-0.9387762546539307,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.48990267515182495,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([4.036538600921631,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6637966632843018,],\n",
+      "[-0.13379549980163574,],\n",
+      "[0.5301522016525269,],\n",
+      "[-0.996950089931488,],\n",
+      "[-0.10863612592220306,],\n",
+      "[0.44324690103530884,],\n",
+      "[-0.5411522388458252,],\n",
+      "[0.8943691849708557,],\n",
+      "[0.8032963275909424,],\n",
+      "[-0.9387762546539307,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.18464016914367676,],\n",
+      "[-0.01737290807068348,],\n",
+      "[-0.1511504054069519,],\n",
+      "[0.03029652126133442,],\n",
+      "[-0.003620665520429611,],\n",
+      "[0.0004926917608827353,],\n",
+      "[-0.13484589755535126,],\n",
+      "[-0.46710723638534546,],\n",
+      "[-0.09121084213256836,],\n",
+      "[-0.004304957576096058,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.48990267515182495,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-4.018233776092529,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6636120080947876,],\n",
+      "[-0.13377812504768372,],\n",
+      "[0.5303033590316772,],\n",
+      "[-0.9969803690910339,],\n",
+      "[-0.10863250494003296,],\n",
+      "[0.44324639439582825,],\n",
+      "[-0.5410174131393433,],\n",
+      "[0.8948363065719604,],\n",
+      "[0.8033875226974487,],\n",
+      "[-0.9387719631195068,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.485884428024292,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.1368095874786377,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6636120080947876,],\n",
+      "[-0.13377812504768372,],\n",
+      "[0.5303033590316772,],\n",
+      "[-0.9969803690910339,],\n",
+      "[-0.10863250494003296,],\n",
+      "[0.44324639439582825,],\n",
+      "[-0.5410174131393433,],\n",
+      "[0.8948363065719604,],\n",
+      "[0.8033875226974487,],\n",
+      "[-0.9387719631195068,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.12360512465238571,],\n",
+      "[-0.012205633334815502,],\n",
+      "[-0.1094035729765892,],\n",
+      "[0.022177211940288544,],\n",
+      "[-0.06886404752731323,],\n",
+      "[-0.015093812718987465,],\n",
+      "[-0.25298309326171875,],\n",
+      "[-0.478471577167511,],\n",
+      "[-0.0973471999168396,],\n",
+      "[-0.004317248705774546,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.485884428024292,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.542208433151245,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6634883880615234,],\n",
+      "[-0.13376592099666595,],\n",
+      "[0.5304127335548401,],\n",
+      "[-0.9970025420188904,],\n",
+      "[-0.10856363922357559,],\n",
+      "[0.4432614743709564,],\n",
+      "[-0.5407644510269165,],\n",
+      "[0.8953147530555725,],\n",
+      "[0.8034848570823669,],\n",
+      "[-0.938767671585083,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4823422133922577,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.872816562652588,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6634883880615234,],\n",
+      "[-0.13376592099666595,],\n",
+      "[0.5304127335548401,],\n",
+      "[-0.9970025420188904,],\n",
+      "[-0.10856363922357559,],\n",
+      "[0.4432614743709564,],\n",
+      "[-0.5407644510269165,],\n",
+      "[0.8953147530555725,],\n",
+      "[0.8034848570823669,],\n",
+      "[-0.938767671585083,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.07208661735057831,],\n",
+      "[-0.007449381984770298,],\n",
+      "[-0.06917523592710495,],\n",
+      "[0.014202384278178215,],\n",
+      "[-0.1775369942188263,],\n",
+      "[-0.04076658934354782,],\n",
+      "[-0.42342108488082886,],\n",
+      "[-0.48892486095428467,],\n",
+      "[-0.1023268848657608,],\n",
+      "[-0.004327971022576094,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4823422133922577,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.7370193004608154,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.66341632604599,],\n",
+      "[-0.13375847041606903,],\n",
+      "[0.5304819345474243,],\n",
+      "[-0.9970167279243469,],\n",
+      "[-0.1083860993385315,],\n",
+      "[0.4433022439479828,],\n",
+      "[-0.5403410196304321,],\n",
+      "[0.895803689956665,],\n",
+      "[0.8035871982574463,],\n",
+      "[-0.9387633204460144,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.47960519790649414,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.070175051689148,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.66341632604599,],\n",
+      "[-0.13375847041606903,],\n",
+      "[0.5304819345474243,],\n",
+      "[-0.9970167279243469,],\n",
+      "[-0.1083860993385315,],\n",
+      "[0.4433022439479828,],\n",
+      "[-0.5403410196304321,],\n",
+      "[0.895803689956665,],\n",
+      "[0.8035871982574463,],\n",
+      "[-0.9387633204460144,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04095807299017906,],\n",
+      "[-0.004402735736221075,],\n",
+      "[-0.04267048463225365,],\n",
+      "[0.008884902112185955,],\n",
+      "[-0.2669855058193207,],\n",
+      "[-0.061598196625709534,],\n",
+      "[-0.5584482550621033,],\n",
+      "[-0.49510619044303894,],\n",
+      "[-0.10524969547986984,],\n",
+      "[-0.004333906807005405,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.47960519790649414,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.0689849853515625,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6633753776550293,],\n",
+      "[-0.13375407457351685,],\n",
+      "[0.5305246114730835,],\n",
+      "[-0.9970256090164185,],\n",
+      "[-0.10811911523342133,],\n",
+      "[0.44336384534835815,],\n",
+      "[-0.5397825837135315,],\n",
+      "[0.8962988257408142,],\n",
+      "[0.8036924600601196,],\n",
+      "[-0.9387589693069458,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4775362014770508,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.925575315952301,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6633753776550293,],\n",
+      "[-0.13375407457351685,],\n",
+      "[0.5305246114730835,],\n",
+      "[-0.9970256090164185,],\n",
+      "[-0.10811911523342133,],\n",
+      "[0.44336384534835815,],\n",
+      "[-0.5397825837135315,],\n",
+      "[0.8962988257408142,],\n",
+      "[0.8036924600601196,],\n",
+      "[-0.9387589693069458,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.028540803119540215,],\n",
+      "[-0.0031756057869642973,],\n",
+      "[-0.03230832889676094,],\n",
+      "[0.006829726975411177,],\n",
+      "[-0.28655630350112915,],\n",
+      "[-0.0659271627664566,],\n",
+      "[-0.5981960296630859,],\n",
+      "[-0.4965459704399109,],\n",
+      "[-0.1064036563038826,],\n",
+      "[-0.004335516132414341,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4775362014770508,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.924135684967041,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6633468270301819,],\n",
+      "[-0.13375090062618256,],\n",
+      "[0.5305569171905518,],\n",
+      "[-0.9970324635505676,],\n",
+      "[-0.10783255845308304,],\n",
+      "[0.44342976808547974,],\n",
+      "[-0.5391843914985657,],\n",
+      "[0.896795392036438,],\n",
+      "[0.8037988543510437,],\n",
+      "[-0.9387546181678772,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.475612074136734,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.407606601715088,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6633468270301819,],\n",
+      "[-0.13375090062618256,],\n",
+      "[0.5305569171905518,],\n",
+      "[-0.9970324635505676,],\n",
+      "[-0.10783255845308304,],\n",
+      "[0.44342976808547974,],\n",
+      "[-0.5391843914985657,],\n",
+      "[0.896795392036438,],\n",
+      "[0.8037988543510437,],\n",
+      "[-0.9387546181678772,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.026303347200155258,],\n",
+      "[-0.0030273119919002056,],\n",
+      "[-0.03238249570131302,],\n",
+      "[0.00695376330986619,],\n",
+      "[-0.22800612449645996,],\n",
+      "[-0.05212101340293884,],\n",
+      "[-0.5442783832550049,],\n",
+      "[-0.4948323965072632,],\n",
+      "[-0.10663972049951553,],\n",
+      "[-0.004334884695708752,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.475612074136734,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.3728513717651367,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6633205413818359,],\n",
+      "[-0.1337478756904602,],\n",
+      "[0.5305892825126648,],\n",
+      "[-0.9970394372940063,],\n",
+      "[-0.10760455578565598,],\n",
+      "[0.4434818923473358,],\n",
+      "[-0.5386401414871216,],\n",
+      "[0.8972902297973633,],\n",
+      "[0.8039054870605469,],\n",
+      "[-0.9387502670288086,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4732392132282257,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.4435455799102783,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6633205413818359,],\n",
+      "[-0.1337478756904602,],\n",
+      "[0.5305892825126648,],\n",
+      "[-0.9970394372940063,],\n",
+      "[-0.10760455578565598,],\n",
+      "[0.4434818923473358,],\n",
+      "[-0.5386401414871216,],\n",
+      "[0.8972902297973633,],\n",
+      "[0.8039054870605469,],\n",
+      "[-0.9387502670288086,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.025841079652309418,],\n",
+      "[-0.0030884514562785625,],\n",
+      "[-0.03463676944375038,],\n",
+      "[0.007556390482932329,],\n",
+      "[-0.13057252764701843,],\n",
+      "[-0.02953118085861206,],\n",
+      "[-0.4522415101528168,],\n",
+      "[-0.4920283854007721,],\n",
+      "[-0.10671732574701309,],\n",
+      "[-0.004334018100053072,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4732392132282257,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.126366376876831,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632946729660034,],\n",
+      "[-0.13374479115009308,],\n",
+      "[0.5306239128112793,],\n",
+      "[-0.9970470070838928,],\n",
+      "[-0.10747398436069489,],\n",
+      "[0.443511426448822,],\n",
+      "[-0.5381879210472107,],\n",
+      "[0.8977822661399841,],\n",
+      "[0.8040121793746948,],\n",
+      "[-0.93874591588974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4701128602027893,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.4438486099243164,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632946729660034,],\n",
+      "[-0.13374479115009308,],\n",
+      "[0.5306239128112793,],\n",
+      "[-0.9970470070838928,],\n",
+      "[-0.10747398436069489,],\n",
+      "[0.443511426448822,],\n",
+      "[-0.5381879210472107,],\n",
+      "[0.8977822661399841,],\n",
+      "[0.8040121793746948,],\n",
+      "[-0.93874591588974,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.022833481431007385,],\n",
+      "[-0.0028506137896329165,],\n",
+      "[-0.033357538282871246,],\n",
+      "[0.007391863036900759,],\n",
+      "[-0.05714303255081177,],\n",
+      "[-0.012724463827908039,],\n",
+      "[-0.3925269544124603,],\n",
+      "[-0.4895280599594116,],\n",
+      "[-0.1069970577955246,],\n",
+      "[-0.004333763383328915,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4701128602027893,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.7115273475646973,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632718443870544,],\n",
+      "[-0.13374194502830505,],\n",
+      "[0.5306572914123535,],\n",
+      "[-0.997054398059845,],\n",
+      "[-0.10741683840751648,],\n",
+      "[0.44352415204048157,],\n",
+      "[-0.53779536485672,],\n",
+      "[0.8982717990875244,],\n",
+      "[0.8041191697120667,],\n",
+      "[-0.9387415647506714,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4664013385772705,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.5297634601593018,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632718443870544,],\n",
+      "[-0.13374194502830505,],\n",
+      "[0.5306572914123535,],\n",
+      "[-0.997054398059845,],\n",
+      "[-0.10741683840751648,],\n",
+      "[0.44352415204048157,],\n",
+      "[-0.53779536485672,],\n",
+      "[0.8982717990875244,],\n",
+      "[0.8041191697120667,],\n",
+      "[-0.9387415647506714,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.01718447357416153,],\n",
+      "[-0.002250076038762927,],\n",
+      "[-0.027381813153624535,],\n",
+      "[0.006161924451589584,],\n",
+      "[-0.05533871799707413,],\n",
+      "[-0.012297373265028,],\n",
+      "[-0.4088374078273773,],\n",
+      "[-0.4877461791038513,],\n",
+      "[-0.10747180879116058,],\n",
+      "[-0.004334136378020048,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4664013385772705,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.757530450820923,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632546782493591,],\n",
+      "[-0.13373969495296478,],\n",
+      "[0.5306846499443054,],\n",
+      "[-0.9970605373382568,],\n",
+      "[-0.10736150294542313,],\n",
+      "[0.4435364603996277,],\n",
+      "[-0.5373865365982056,],\n",
+      "[0.8987595438957214,],\n",
+      "[0.8042266368865967,],\n",
+      "[-0.9387372136116028,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4626438021659851,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.5933687686920166,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632546782493591,],\n",
+      "[-0.13373969495296478,],\n",
+      "[0.5306846499443054,],\n",
+      "[-0.9970605373382568,],\n",
+      "[-0.10736150294542313,],\n",
+      "[0.4435364603996277,],\n",
+      "[-0.5373865365982056,],\n",
+      "[0.8987595438957214,],\n",
+      "[0.8042266368865967,],\n",
+      "[-0.9387372136116028,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.010942076332867146,],\n",
+      "[-0.001502196304500103,],\n",
+      "[-0.019018465653061867,],\n",
+      "[0.004346742294728756,],\n",
+      "[-0.1281520426273346,],\n",
+      "[-0.02847146987915039,],\n",
+      "[-0.49680113792419434,],\n",
+      "[-0.48648080229759216,],\n",
+      "[-0.10797197371721268,],\n",
+      "[-0.004334757570177317,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4626438021659851,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.2207906246185303,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632437109947205,],\n",
+      "[-0.1337381899356842,],\n",
+      "[0.5307036638259888,],\n",
+      "[-0.9970648884773254,],\n",
+      "[-0.10723335295915604,],\n",
+      "[0.44356492161750793,],\n",
+      "[-0.5368897318840027,],\n",
+      "[0.8992460370063782,],\n",
+      "[0.8043345808982849,],\n",
+      "[-0.9387328624725342,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4594230055809021,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.4533605575561523,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632437109947205,],\n",
+      "[-0.1337381899356842,],\n",
+      "[0.5307036638259888,],\n",
+      "[-0.9970648884773254,],\n",
+      "[-0.10723335295915604,],\n",
+      "[0.44356492161750793,],\n",
+      "[-0.5368897318840027,],\n",
+      "[0.8992460370063782,],\n",
+      "[0.8043345808982849,],\n",
+      "[-0.9387328624725342,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.006082792766392231,],\n",
+      "[-0.0008716884767636657,],\n",
+      "[-0.01152732688933611,],\n",
+      "[0.002677040873095393,],\n",
+      "[-0.23342004418373108,],\n",
+      "[-0.051590222865343094,],\n",
+      "[-0.6104710102081299,],\n",
+      "[-0.48539257049560547,],\n",
+      "[-0.10834797471761703,],\n",
+      "[-0.004335233010351658,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4594230055809021,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.4111056327819824,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632376313209534,],\n",
+      "[-0.13373732566833496,],\n",
+      "[0.5307151675224304,],\n",
+      "[-0.9970675706863403,],\n",
+      "[-0.10699993371963501,],\n",
+      "[0.44361650943756104,],\n",
+      "[-0.5362792611122131,],\n",
+      "[0.899731457233429,],\n",
+      "[0.8044429421424866,],\n",
+      "[-0.9387285113334656,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4570119082927704,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8132330179214478,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632376313209534,],\n",
+      "[-0.13373732566833496,],\n",
+      "[0.5307151675224304,],\n",
+      "[-0.9970675706863403,],\n",
+      "[-0.10699993371963501,],\n",
+      "[0.44361650943756104,],\n",
+      "[-0.5362792611122131,],\n",
+      "[0.899731457233429,],\n",
+      "[0.8044429421424866,],\n",
+      "[-0.9387285113334656,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0033782082609832287,],\n",
+      "[-0.0005022762343287468,],\n",
+      "[-0.006975035183131695,],\n",
+      "[0.0016470405971631408,],\n",
+      "[-0.311123251914978,],\n",
+      "[-0.0684593990445137,],\n",
+      "[-0.6921523213386536,],\n",
+      "[-0.4842424690723419,],\n",
+      "[-0.10855168104171753,],\n",
+      "[-0.004335511010140181,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4570119082927704,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8035889863967896,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632342338562012,],\n",
+      "[-0.13373681902885437,],\n",
+      "[0.5307221412658691,],\n",
+      "[-0.997069239616394,],\n",
+      "[-0.10668881237506866,],\n",
+      "[0.44368496537208557,],\n",
+      "[-0.5355871319770813,],\n",
+      "[0.9002156853675842,],\n",
+      "[0.8045514822006226,],\n",
+      "[-0.938724160194397,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.45520833134651184,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7657997012138367,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632342338562012,],\n",
+      "[-0.13373681902885437,],\n",
+      "[0.5307221412658691,],\n",
+      "[-0.997069239616394,],\n",
+      "[-0.10668881237506866,],\n",
+      "[0.44368496537208557,],\n",
+      "[-0.5355871319770813,],\n",
+      "[0.9002156853675842,],\n",
+      "[0.8045514822006226,],\n",
+      "[-0.938724160194397,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.002433427609503269,],\n",
+      "[-0.00037373974919319153,],\n",
+      "[-0.005471008829772472,],\n",
+      "[0.0013142626266926527,],\n",
+      "[-0.3197171688079834,],\n",
+      "[-0.07017218321561813,],\n",
+      "[-0.7081785798072815,],\n",
+      "[-0.4828658401966095,],\n",
+      "[-0.10862244665622711,],\n",
+      "[-0.004335577134042978,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.45520833134651184,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7501976490020752,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632317900657654,],\n",
+      "[-0.13373644649982452,],\n",
+      "[0.5307276248931885,],\n",
+      "[-0.9970705509185791,],\n",
+      "[-0.10636909306049347,],\n",
+      "[0.4437551498413086,],\n",
+      "[-0.5348789691925049,],\n",
+      "[0.9006985425949097,],\n",
+      "[0.8046600818634033,],\n",
+      "[-0.9387198090553284,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.45345813035964966,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2838892936706543,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632317900657654,],\n",
+      "[-0.13373644649982452,],\n",
+      "[0.5307276248931885,],\n",
+      "[-0.9970705509185791,],\n",
+      "[-0.10636909306049347,],\n",
+      "[0.4437551498413086,],\n",
+      "[-0.5348789691925049,],\n",
+      "[0.9006985425949097,],\n",
+      "[0.8046600818634033,],\n",
+      "[-0.9387198090553284,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0023383742664009333,],\n",
+      "[-0.00037080611218698323,],\n",
+      "[-0.0057242028415203094,],\n",
+      "[0.0013989969156682491,],\n",
+      "[-0.2602478861808777,],\n",
+      "[-0.057133518159389496,],\n",
+      "[-0.6667798757553101,],\n",
+      "[-0.4810798168182373,],\n",
+      "[-0.10863204300403595,],\n",
+      "[-0.004335497040301561,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.45345813035964966,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.266176700592041,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632294654846191,],\n",
+      "[-0.13373607397079468,],\n",
+      "[0.5307333469390869,],\n",
+      "[-0.9970719218254089,],\n",
+      "[-0.10610884428024292,],\n",
+      "[0.4438122808933258,],\n",
+      "[-0.5342121720314026,],\n",
+      "[0.9011796116828918,],\n",
+      "[0.8047687411308289,],\n",
+      "[-0.9387154579162598,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4511919617652893,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.2813260555267334,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632294654846191,],\n",
+      "[-0.13373607397079468,],\n",
+      "[0.5307333469390869,],\n",
+      "[-0.9970719218254089,],\n",
+      "[-0.10610884428024292,],\n",
+      "[0.4438122808933258,],\n",
+      "[-0.5342121720314026,],\n",
+      "[0.9011796116828918,],\n",
+      "[0.8047687411308289,],\n",
+      "[-0.9387154579162598,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0023128073662519455,],\n",
+      "[-0.00037990466807968915,],\n",
+      "[-0.006166981533169746,],\n",
+      "[0.0015327560249716043,],\n",
+      "[-0.1755070835351944,],\n",
+      "[-0.03864318132400513,],\n",
+      "[-0.6074382662773132,],\n",
+      "[-0.4787391424179077,],\n",
+      "[-0.1086357831954956,],\n",
+      "[-0.004335472825914621,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4511919617652893,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.020811080932617,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632271409034729,],\n",
+      "[-0.13373570144176483,],\n",
+      "[0.5307394862174988,],\n",
+      "[-0.9970734715461731,],\n",
+      "[-0.10593333840370178,],\n",
+      "[0.44385093450546265,],\n",
+      "[-0.5336047410964966,],\n",
+      "[0.9016583561897278,],\n",
+      "[0.8048774003982544,],\n",
+      "[-0.9387111067771912,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.44817113876342773,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.129657030105591,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632271409034729,],\n",
+      "[-0.13373570144176483,],\n",
+      "[0.5307394862174988,],\n",
+      "[-0.9970734715461731,],\n",
+      "[-0.10593333840370178,],\n",
+      "[0.44385093450546265,],\n",
+      "[-0.5336047410964966,],\n",
+      "[0.9016583561897278,],\n",
+      "[0.8048774003982544,],\n",
+      "[-0.9387111067771912,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0020096185617148876,],\n",
+      "[-0.0003435338439885527,],\n",
+      "[-0.005840560887008905,],\n",
+      "[0.0014752420829609036,],\n",
+      "[-0.12349852919578552,],\n",
+      "[-0.027017368003726006,],\n",
+      "[-0.5739790201187134,],\n",
+      "[-0.47583475708961487,],\n",
+      "[-0.10865820199251175,],\n",
+      "[-0.004335454665124416,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.44817113876342773,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.5381622314453125,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632251143455505,],\n",
+      "[-0.13373535871505737,],\n",
+      "[0.5307453274726868,],\n",
+      "[-0.9970749616622925,],\n",
+      "[-0.10580983757972717,],\n",
+      "[0.4438779652118683,],\n",
+      "[-0.5330307483673096,],\n",
+      "[0.9021341800689697,],\n",
+      "[0.8049860596656799,],\n",
+      "[-0.9387067556381226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.44463297724723816,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.0379061698913574,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632251143455505,],\n",
+      "[-0.13373535871505737,],\n",
+      "[0.5307453274726868,],\n",
+      "[-0.9970749616622925,],\n",
+      "[-0.10580983757972717,],\n",
+      "[0.4438779652118683,],\n",
+      "[-0.5330307483673096,],\n",
+      "[0.9021341800689697,],\n",
+      "[0.8049860596656799,],\n",
+      "[-0.9387067556381226,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0014686838258057833,],\n",
+      "[-0.0002620294690132141,],\n",
+      "[-0.004654419608414173,],\n",
+      "[0.001194189302623272,],\n",
+      "[-0.14022132754325867,],\n",
+      "[-0.02981869876384735,],\n",
+      "[-0.5903998017311096,],\n",
+      "[-0.4726124703884125,],\n",
+      "[-0.10869500041007996,],\n",
+      "[-0.004335495643317699,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.44463297724723816,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.4859185218811035,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632236242294312,],\n",
+      "[-0.13373509049415588,],\n",
+      "[0.5307499766349792,],\n",
+      "[-0.997076153755188,],\n",
+      "[-0.10566961765289307,],\n",
+      "[0.44390779733657837,],\n",
+      "[-0.5324403643608093,],\n",
+      "[0.9026067852973938,],\n",
+      "[0.8050947785377502,],\n",
+      "[-0.938702404499054,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4411470592021942,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.0692856311798096,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632236242294312,],\n",
+      "[-0.13373509049415588,],\n",
+      "[0.5307499766349792,],\n",
+      "[-0.997076153755188,],\n",
+      "[-0.10566961765289307,],\n",
+      "[0.44390779733657837,],\n",
+      "[-0.5324403643608093,],\n",
+      "[0.9026067852973938,],\n",
+      "[0.8050947785377502,],\n",
+      "[-0.938702404499054,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0008991042268462479,],\n",
+      "[-0.00016727729234844446,],\n",
+      "[-0.0031067505478858948,],\n",
+      "[0.0008097035461105406,],\n",
+      "[-0.21756111085414886,],\n",
+      "[-0.04547234624624252,],\n",
+      "[-0.6499828100204468,],\n",
+      "[-0.46951034665107727,],\n",
+      "[-0.10873214155435562,],\n",
+      "[-0.004335506819188595,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4411470592021942,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.8769986629486084,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632227301597595,],\n",
+      "[-0.13373492658138275,],\n",
+      "[0.5307530760765076,],\n",
+      "[-0.9970769882202148,],\n",
+      "[-0.1054520532488823,],\n",
+      "[0.443953275680542,],\n",
+      "[-0.5317903757095337,],\n",
+      "[0.9030762910842896,],\n",
+      "[0.8052034974098206,],\n",
+      "[-0.9386980533599854,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.43827006220817566,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.0626503229141235,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632227301597595,],\n",
+      "[-0.13373492658138275,],\n",
+      "[0.5307530760765076,],\n",
+      "[-0.9970769882202148,],\n",
+      "[-0.1054520532488823,],\n",
+      "[0.443953275680542,],\n",
+      "[-0.5317903757095337,],\n",
+      "[0.9030762910842896,],\n",
+      "[0.8052034974098206,],\n",
+      "[-0.9386980533599854,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00047791923861950636,],\n",
+      "[-9.23397092265077e-05,],\n",
+      "[-0.0017994983354583383,],\n",
+      "[0.00047667144099250436,],\n",
+      "[-0.31175053119659424,],\n",
+      "[-0.06493367999792099,],\n",
+      "[-0.7210558652877808,],\n",
+      "[-0.46694183349609375,],\n",
+      "[-0.10875771194696426,],\n",
+      "[-0.004335533361881971,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.43827006220817566,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.0616986751556396,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632222533226013,],\n",
+      "[-0.1337348371744156,],\n",
+      "[0.5307548642158508,],\n",
+      "[-0.997077465057373,],\n",
+      "[-0.1051403060555458,],\n",
+      "[0.4440182149410248,],\n",
+      "[-0.5310693383216858,],\n",
+      "[0.90354323387146,],\n",
+      "[0.8053122758865356,],\n",
+      "[-0.9386937022209167,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4362083673477173,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5739565491676331,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632222533226013,],\n",
+      "[-0.1337348371744156,],\n",
+      "[0.5307548642158508,],\n",
+      "[-0.997077465057373,],\n",
+      "[-0.1051403060555458,],\n",
+      "[0.4440182149410248,],\n",
+      "[-0.5310693383216858,],\n",
+      "[0.90354323387146,],\n",
+      "[0.8053122758865356,],\n",
+      "[-0.9386937022209167,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0002605319896247238,],\n",
+      "[-5.20071953360457e-05,],\n",
+      "[-0.0010681095300242305,],\n",
+      "[0.0002877785882446915,],\n",
+      "[-0.37277117371559143,],\n",
+      "[-0.07756871730089188,],\n",
+      "[-0.7672025561332703,],\n",
+      "[-0.4650508761405945,],\n",
+      "[-0.10877160727977753,],\n",
+      "[-0.004335570149123669,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4362083673477173,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.515197992324829,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632220149040222,],\n",
+      "[-0.133734792470932,],\n",
+      "[0.5307559370994568,],\n",
+      "[-0.9970777630805969,],\n",
+      "[-0.10476753115653992,],\n",
+      "[0.44409579038619995,],\n",
+      "[-0.530302107334137,],\n",
+      "[0.9040082693099976,],\n",
+      "[0.8054210543632507,],\n",
+      "[-0.9386893510818481,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4346931576728821,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6002265810966492,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632220149040222,],\n",
+      "[-0.133734792470932,],\n",
+      "[0.5307559370994568,],\n",
+      "[-0.9970777630805969,],\n",
+      "[-0.10476753115653992,],\n",
+      "[0.44409579038619995,],\n",
+      "[-0.530302107334137,],\n",
+      "[0.9040082693099976,],\n",
+      "[0.8054210543632507,],\n",
+      "[-0.9386893510818481,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00019762372539844364,],\n",
+      "[-4.0620547224534675e-05,],\n",
+      "[-0.0008816892514005303,],\n",
+      "[0.00024174251302611083,],\n",
+      "[-0.3731462359428406,],\n",
+      "[-0.07734764367341995,],\n",
+      "[-0.7702398300170898,],\n",
+      "[-0.4635808765888214,],\n",
+      "[-0.10877519100904465,],\n",
+      "[-0.004335565026849508,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4346931576728821,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.549485206604004,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632218360900879,],\n",
+      "[-0.13373474776744843,],\n",
+      "[0.5307568311691284,],\n",
+      "[-0.997078001499176,],\n",
+      "[-0.10439438372850418,],\n",
+      "[0.444173127412796,],\n",
+      "[-0.5295318961143494,],\n",
+      "[0.9044718742370605,],\n",
+      "[0.8055298328399658,],\n",
+      "[-0.9386849999427795,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.433143675327301,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.128089427947998,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632218360900879,],\n",
+      "[-0.13373474776744843,],\n",
+      "[0.5307568311691284,],\n",
+      "[-0.997078001499176,],\n",
+      "[-0.10439438372850418,],\n",
+      "[0.444173127412796,],\n",
+      "[-0.5295318961143494,],\n",
+      "[0.9044718742370605,],\n",
+      "[0.8055298328399658,],\n",
+      "[-0.9386849999427795,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0002009579329751432,],\n",
+      "[-4.25368343712762e-05,],\n",
+      "[-0.0009756840881891549,],\n",
+      "[0.00027222264907322824,],\n",
+      "[-0.32227689027786255,],\n",
+      "[-0.06608667224645615,],\n",
+      "[-0.7391725778579712,],\n",
+      "[-0.46201691031455994,],\n",
+      "[-0.10877508670091629,],\n",
+      "[-0.004335561767220497,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.433143675327301,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1242289543151855,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632216572761536,],\n",
+      "[-0.13373470306396484,],\n",
+      "[0.5307577848434448,],\n",
+      "[-0.9970782995223999,],\n",
+      "[-0.10407210886478424,],\n",
+      "[0.44423922896385193,],\n",
+      "[-0.5287927389144897,],\n",
+      "[0.9049338698387146,],\n",
+      "[0.8056386113166809,],\n",
+      "[-0.9386806488037109,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.43101945519447327,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.0611889362335205,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632216572761536,],\n",
+      "[-0.13373470306396484,],\n",
+      "[0.5307577848434448,],\n",
+      "[-0.9970782995223999,],\n",
+      "[-0.10407210886478424,],\n",
+      "[0.44423922896385193,],\n",
+      "[-0.5287927389144897,],\n",
+      "[0.9049338698387146,],\n",
+      "[0.8056386113166809,],\n",
+      "[-0.9386806488037109,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0002014814381254837,],\n",
+      "[-4.406089647090994e-05,],\n",
+      "[-0.0010651237098500133,],\n",
+      "[0.00030223483918234706,],\n",
+      "[-0.2603999376296997,],\n",
+      "[-0.05211314931511879,],\n",
+      "[-0.7007242441177368,],\n",
+      "[-0.45987823605537415,],\n",
+      "[-0.10877498239278793,],\n",
+      "[-0.004335558507591486,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.43101945519447327,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.8713700771331787,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632214784622192,],\n",
+      "[-0.13373465836048126,],\n",
+      "[0.5307588577270508,],\n",
+      "[-0.9970785975456238,],\n",
+      "[-0.10381171107292175,],\n",
+      "[0.444291353225708,],\n",
+      "[-0.5280920267105103,],\n",
+      "[0.9053937196731567,],\n",
+      "[0.805747389793396,],\n",
+      "[-0.9386762976646423,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.428148090839386,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.7530930042266846,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632214784622192,],\n",
+      "[-0.13373465836048126,],\n",
+      "[0.5307588577270508,],\n",
+      "[-0.9970785975456238,],\n",
+      "[-0.10381171107292175,],\n",
+      "[0.444291353225708,],\n",
+      "[-0.5280920267105103,],\n",
+      "[0.9053937196731567,],\n",
+      "[0.805747389793396,],\n",
+      "[-0.9386762976646423,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00017271183605771512,],\n",
+      "[-3.9180671592475846e-05,],\n",
+      "[-0.0009948444785550237,],\n",
+      "[0.0002868863521143794,],\n",
+      "[-0.23283401131629944,],\n",
+      "[-0.04494169354438782,],\n",
+      "[-0.6817854642868042,],\n",
+      "[-0.4570252299308777,],\n",
+      "[-0.10877712070941925,],\n",
+      "[-0.004335555713623762,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.428148090839386,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.318493366241455,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632212996482849,],\n",
+      "[-0.13373461365699768,],\n",
+      "[0.530759871006012,],\n",
+      "[-0.9970788955688477,],\n",
+      "[-0.10357888042926788,],\n",
+      "[0.44433629512786865,],\n",
+      "[-0.5274102687835693,],\n",
+      "[0.9058507680892944,],\n",
+      "[0.8058561682701111,],\n",
+      "[-0.9386719465255737,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4248296022415161,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.5207834243774414,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632212996482849,],\n",
+      "[-0.13373461365699768,],\n",
+      "[0.530759871006012,],\n",
+      "[-0.9970788955688477,],\n",
+      "[-0.10357888042926788,],\n",
+      "[0.44433629512786865,],\n",
+      "[-0.5274102687835693,],\n",
+      "[0.9058507680892944,],\n",
+      "[0.8058561682701111,],\n",
+      "[-0.9386719465255737,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0001225775049533695,],\n",
+      "[-2.891064104915131e-05,],\n",
+      "[-0.000769656733609736,],\n",
+      "[0.0002254590071970597,],\n",
+      "[-0.259912371635437,],\n",
+      "[-0.049322888255119324,],\n",
+      "[-0.6952139735221863,],\n",
+      "[-0.4537578523159027,],\n",
+      "[-0.10878001898527145,],\n",
+      "[-0.004335553385317326,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4248296022415161,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.1753997802734375,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632211804389954,],\n",
+      "[-0.1337345838546753,],\n",
+      "[0.530760645866394,],\n",
+      "[-0.9970791339874268,],\n",
+      "[-0.1033189669251442,],\n",
+      "[0.4443856179714203,],\n",
+      "[-0.5267150402069092,],\n",
+      "[0.9063045382499695,],\n",
+      "[0.8059649467468262,],\n",
+      "[-0.9386675953865051,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.42165419459342957,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.5749974250793457,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632211804389954,],\n",
+      "[-0.1337345838546753,],\n",
+      "[0.530760645866394,],\n",
+      "[-0.9970791339874268,],\n",
+      "[-0.1033189669251442,],\n",
+      "[0.4443856179714203,],\n",
+      "[-0.5267150402069092,],\n",
+      "[0.9063045382499695,],\n",
+      "[0.8059649467468262,],\n",
+      "[-0.9386675953865051,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.186464063124731e-05,],\n",
+      "[-1.7602242223802023e-05,],\n",
+      "[-0.0004918001359328628,],\n",
+      "[0.0001463637308916077,],\n",
+      "[-0.3262772262096405,],\n",
+      "[-0.06272263079881668,],\n",
+      "[-0.7342639565467834,],\n",
+      "[-0.45063942670822144,],\n",
+      "[-0.10878219455480576,],\n",
+      "[-0.004335551988333464,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.42165419459342957,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.509977340698242,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632211208343506,],\n",
+      "[-0.1337345689535141,],\n",
+      "[0.5307611227035522,],\n",
+      "[-0.9970792531967163,],\n",
+      "[-0.10299269109964371,],\n",
+      "[0.4444483518600464,],\n",
+      "[-0.5259807705879211,],\n",
+      "[0.9067551493644714,],\n",
+      "[0.8060737252235413,],\n",
+      "[-0.9386632442474365,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4191442131996155,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7283968329429626,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632211208343506,],\n",
+      "[-0.1337345689535141,],\n",
+      "[0.5307611227035522,],\n",
+      "[-0.9970792531967163,],\n",
+      "[-0.10299269109964371,],\n",
+      "[0.4444483518600464,],\n",
+      "[-0.5259807705879211,],\n",
+      "[0.9067551493644714,],\n",
+      "[0.8060737252235413,],\n",
+      "[-0.9386632442474365,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.6248809919925407e-05,],\n",
+      "[-9.18474506761413e-06,],\n",
+      "[-0.0002701895427890122,],\n",
+      "[8.174658432835713e-05,],\n",
+      "[-0.3956589102745056,],\n",
+      "[-0.077540323138237,],\n",
+      "[-0.7773784399032593,],\n",
+      "[-0.4481743574142456,],\n",
+      "[-0.10878439992666245,],\n",
+      "[-0.004335551057010889,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4191442131996155,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7069222927093506,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307614207267761,],\n",
+      "[-0.9970793128013611,],\n",
+      "[-0.10259703546762466,],\n",
+      "[0.44452589750289917,],\n",
+      "[-0.5252034068107605,],\n",
+      "[0.9072033166885376,],\n",
+      "[0.8061825037002563,],\n",
+      "[-0.9386588931083679,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.41743728518486023,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3810309171676636,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307614207267761,],\n",
+      "[-0.9970793128013611,],\n",
+      "[-0.10259703546762466,],\n",
+      "[0.44452589750289917,],\n",
+      "[-0.5252034068107605,],\n",
+      "[0.9072033166885376,],\n",
+      "[0.8061825037002563,],\n",
+      "[-0.9386588931083679,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.944572795764543e-05,],\n",
+      "[-5.074372438684804e-06,],\n",
+      "[-0.0001577446673763916,],\n",
+      "[4.855753650190309e-05,],\n",
+      "[-0.43456438183784485,],\n",
+      "[-0.08597969263792038,],\n",
+      "[-0.8020826578140259,],\n",
+      "[-0.4464884102344513,],\n",
+      "[-0.10878513008356094,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.41743728518486023,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.2345542907714844,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307615995407104,],\n",
+      "[-0.9970793724060059,],\n",
+      "[-0.10216247290372849,],\n",
+      "[0.44461187720298767,],\n",
+      "[-0.5244013071060181,],\n",
+      "[0.90764981508255,],\n",
+      "[0.8062912821769714,],\n",
+      "[-0.9386545419692993,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.41620272397994995,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.46308910846710205,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307615995407104,],\n",
+      "[-0.9970793724060059,],\n",
+      "[-0.10216247290372849,],\n",
+      "[0.44461187720298767,],\n",
+      "[-0.5244013071060181,],\n",
+      "[0.90764981508255,],\n",
+      "[0.8062912821769714,],\n",
+      "[-0.9386545419692993,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.5900579455774277e-05,],\n",
+      "[-4.262170023139333e-06,],\n",
+      "[-0.00014031196769792587,],\n",
+      "[4.396398435346782e-05,],\n",
+      "[-0.42937320470809937,],\n",
+      "[-0.08445799350738525,],\n",
+      "[-0.7991408705711365,],\n",
+      "[-0.44525688886642456,],\n",
+      "[-0.10878512263298035,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.41620272397994995,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.3610126972198486,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.53076171875,],\n",
+      "[-0.9970794320106506,],\n",
+      "[-0.10173310339450836,],\n",
+      "[0.4446963369846344,],\n",
+      "[-0.5236021876335144,],\n",
+      "[0.9080950617790222,],\n",
+      "[0.8064000606536865,],\n",
+      "[-0.9386501908302307,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4148417115211487,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.9917455911636353,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.53076171875,],\n",
+      "[-0.9970794320106506,],\n",
+      "[-0.10173310339450836,],\n",
+      "[0.4446963369846344,],\n",
+      "[-0.5236021876335144,],\n",
+      "[0.9080950617790222,],\n",
+      "[0.8064000606536865,],\n",
+      "[-0.9386501908302307,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.7258997104363516e-05,],\n",
+      "[-4.755157988256542e-06,],\n",
+      "[-0.00016569194849580526,],\n",
+      "[5.2838360716123134e-05,],\n",
+      "[-0.39227092266082764,],\n",
+      "[-0.07540114223957062,],\n",
+      "[-0.7764885425567627,],\n",
+      "[-0.4438929557800293,],\n",
+      "[-0.10878511518239975,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4148417115211487,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.9917278289794922,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307618975639343,],\n",
+      "[-0.9970794916152954,],\n",
+      "[-0.10134083032608032,],\n",
+      "[0.44477173686027527,],\n",
+      "[-0.5228257179260254,],\n",
+      "[0.9085389375686646,],\n",
+      "[0.8065088391304016,],\n",
+      "[-0.9386458396911621,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4128499925136566,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.858839511871338,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307618975639343,],\n",
+      "[-0.9970794916152954,],\n",
+      "[-0.10134083032608032,],\n",
+      "[0.44477173686027527,],\n",
+      "[-0.5228257179260254,],\n",
+      "[0.9085389375686646,],\n",
+      "[0.8065088391304016,],\n",
+      "[-0.9386458396911621,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.7525433577247895e-05,],\n",
+      "[-4.979164714313811e-06,],\n",
+      "[-0.0001831587724154815,],\n",
+      "[5.940847040619701e-05,],\n",
+      "[-0.35370922088623047,],\n",
+      "[-0.06553512811660767,],\n",
+      "[-0.7517712712287903,],\n",
+      "[-0.44189831614494324,],\n",
+      "[-0.10878510773181915,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4128499925136566,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.726785182952881,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307620763778687,],\n",
+      "[-0.9970795512199402,],\n",
+      "[-0.10098712146282196,],\n",
+      "[0.4448372721672058,],\n",
+      "[-0.5220739245414734,],\n",
+      "[0.9089808464050293,],\n",
+      "[0.8066176176071167,],\n",
+      "[-0.9386414885520935,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.41012319922447205,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.407898187637329,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307620763778687,],\n",
+      "[-0.9970795512199402,],\n",
+      "[-0.10098712146282196,],\n",
+      "[0.4448372721672058,],\n",
+      "[-0.5220739245414734,],\n",
+      "[0.9089808464050293,],\n",
+      "[0.8066176176071167,],\n",
+      "[-0.9386414885520935,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.47944820128032e-05,],\n",
+      "[-4.350295512267621e-06,],\n",
+      "[-0.00016843988851178437,],\n",
+      "[5.552903894567862e-05,],\n",
+      "[-0.3424972891807556,],\n",
+      "[-0.061581552028656006,],\n",
+      "[-0.7412003874778748,],\n",
+      "[-0.4391745924949646,],\n",
+      "[-0.10878510028123856,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.41012319922447205,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.103477954864502,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.530762255191803,],\n",
+      "[-0.997079610824585,],\n",
+      "[-0.10064462572336197,],\n",
+      "[0.4448988437652588,],\n",
+      "[-0.5213327407836914,],\n",
+      "[0.9094200134277344,],\n",
+      "[0.8067263960838318,],\n",
+      "[-0.9386371374130249,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4070197343826294,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.069087266921997,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.530762255191803,],\n",
+      "[-0.997079610824585,],\n",
+      "[-0.10064462572336197,],\n",
+      "[0.4448988437652588,],\n",
+      "[-0.5213327407836914,],\n",
+      "[0.9094200134277344,],\n",
+      "[0.8067263960838318,],\n",
+      "[-0.9386371374130249,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.017193790175952e-05,],\n",
+      "[-3.1009517442726064e-06,],\n",
+      "[-0.00012621007044799626,],\n",
+      "[4.2272924474673346e-05,],\n",
+      "[-0.36690062284469604,],\n",
+      "[-0.0661284402012825,],\n",
+      "[-0.7512792944908142,],\n",
+      "[-0.43608012795448303,],\n",
+      "[-0.10878509283065796,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4070197343826294,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.8768606185913086,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307623744010925,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.10027772188186646,],\n",
+      "[0.4449649751186371,],\n",
+      "[-0.5205814838409424,],\n",
+      "[0.9098560810089111,],\n",
+      "[0.8068351745605469,],\n",
+      "[-0.9386327862739563,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4041428864002228,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1751405000686646,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307623744010925,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.10027772188186646,],\n",
+      "[0.4449649751186371,],\n",
+      "[-0.5205814838409424,],\n",
+      "[0.9098560810089111,],\n",
+      "[0.8068351745605469,],\n",
+      "[-0.9386327862739563,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[5.685808901034761e-06,],\n",
+      "[-1.7945718582268455e-06,],\n",
+      "[-7.68649551901035e-05,],\n",
+      "[2.6163068469031714e-05,],\n",
+      "[-0.4125133752822876,],\n",
+      "[-0.076473668217659,],\n",
+      "[-0.7760826349258423,],\n",
+      "[-0.4332122504711151,],\n",
+      "[-0.10878583043813705,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4041428864002228,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1680805683135986,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624340057373,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09986520558595657,],\n",
+      "[0.4450414478778839,],\n",
+      "[-0.519805371761322,],\n",
+      "[0.9102892875671387,],\n",
+      "[0.806943953037262,],\n",
+      "[-0.9386284351348877,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.40197479724884033,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4836312234401703,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624340057373,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09986520558595657,],\n",
+      "[0.4450414478778839,],\n",
+      "[-0.519805371761322,],\n",
+      "[0.9102892875671387,],\n",
+      "[0.806943953037262,],\n",
+      "[-0.9386284351348877,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.705436600081157e-06,],\n",
+      "[-8.808054872133653e-07,],\n",
+      "[-3.982109046773985e-05,],\n",
+      "[1.3783913345832843e-05,],\n",
+      "[-0.45520520210266113,],\n",
+      "[-0.0868091955780983,],\n",
+      "[-0.8013225793838501,],\n",
+      "[-0.4310501515865326,],\n",
+      "[-0.10878583043813705,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.40197479724884033,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.3908735513687134,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09940999746322632,],\n",
+      "[0.44512826204299927,],\n",
+      "[-0.5190040469169617,],\n",
+      "[0.9107203483581543,],\n",
+      "[0.807052731513977,],\n",
+      "[-0.9386240839958191,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.4005839228630066,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.250404953956604,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09940999746322632,],\n",
+      "[0.44512826204299927,],\n",
+      "[-0.5190040469169617,],\n",
+      "[0.9107203483581543,],\n",
+      "[0.807052731513977,],\n",
+      "[-0.9386240839958191,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.443894575459126e-06,],\n",
+      "[-4.83006999729696e-07,],\n",
+      "[-2.3121523554436862e-05,],\n",
+      "[8.14530358184129e-06,],\n",
+      "[-0.4757435917854309,],\n",
+      "[-0.0918581560254097,],\n",
+      "[-0.8136268258094788,],\n",
+      "[-0.42966228723526,],\n",
+      "[-0.10878583043813705,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.4005839228630066,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0008097887039185,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09893425554037094,],\n",
+      "[0.44522011280059814,],\n",
+      "[-0.5181904435157776,],\n",
+      "[0.9111500382423401,],\n",
+      "[0.8071615099906921,],\n",
+      "[-0.9386197328567505,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.399583101272583,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3715556859970093,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09893425554037094,],\n",
+      "[0.44522011280059814,],\n",
+      "[-0.5181904435157776,],\n",
+      "[0.9111500382423401,],\n",
+      "[0.8071615099906921,],\n",
+      "[-0.9386197328567505,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.3045458899796358e-06,],\n",
+      "[-4.475162711514713e-07,],\n",
+      "[-2.2718264517607167e-05,],\n",
+      "[8.148335837177001e-06,],\n",
+      "[-0.468670129776001,],\n",
+      "[-0.08977420628070831,],\n",
+      "[-0.8087864518165588,],\n",
+      "[-0.4286614954471588,],\n",
+      "[-0.10878583043813705,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.399583101272583,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.2191067934036255,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09846558421850204,],\n",
+      "[0.4453098773956299,],\n",
+      "[-0.5173816680908203,],\n",
+      "[0.9115787148475647,],\n",
+      "[0.8072702884674072,],\n",
+      "[-0.9386153817176819,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.39836400747299194,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.9010468125343323,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09846558421850204,],\n",
+      "[0.4453098773956299,],\n",
+      "[-0.5173816680908203,],\n",
+      "[0.9115787148475647,],\n",
+      "[0.8072702884674072,],\n",
+      "[-0.9386153817176819,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.506796706962632e-06,],\n",
+      "[-5.306195589582785e-07,],\n",
+      "[-2.8542996005853638e-05,],\n",
+      "[1.0420777471153997e-05,],\n",
+      "[-0.4442673325538635,],\n",
+      "[-0.0829576626420021,],\n",
+      "[-0.7929679155349731,],\n",
+      "[-0.42744240164756775,],\n",
+      "[-0.10878583043813705,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.39836400747299194,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.898468255996704,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09802131354808807,],\n",
+      "[0.4453928470611572,],\n",
+      "[-0.5165886878967285,],\n",
+      "[0.912006139755249,],\n",
+      "[0.8073790669441223,],\n",
+      "[-0.9386110305786133,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.396465539932251,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.706151008605957,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307624936103821,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09802131354808807,],\n",
+      "[0.4453928470611572,],\n",
+      "[-0.5165886878967285,],\n",
+      "[0.912006139755249,],\n",
+      "[0.8073790669441223,],\n",
+      "[-0.9386110305786133,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.5378866464743624e-06,],\n",
+      "[-5.577424531111319e-07,],\n",
+      "[-3.170934724039398e-05,],\n",
+      "[1.1776144674513489e-05,],\n",
+      "[-0.4221839904785156,],\n",
+      "[-0.07640326023101807,],\n",
+      "[-0.7774237394332886,],\n",
+      "[-0.425540953874588,],\n",
+      "[-0.10878583043813705,],\n",
+      "[-0.004335550591349602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.396465539932251,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.6123976707458496,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307625532150269,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09759912639856339,],\n",
+      "[0.4454692602157593,],\n",
+      "[-0.5158112645149231,],\n",
+      "[0.9124316573143005,],\n",
+      "[0.8074878454208374,],\n",
+      "[-0.9386066794395447,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3938531279563904,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [1/10], Loss: 1.7062\n",
+      "loss_antes tensor([1.7912184000015259,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6632210612297058,],\n",
+      "[-0.1337345540523529,],\n",
+      "[0.5307625532150269,],\n",
+      "[-0.9970796704292297,],\n",
+      "[-0.09759912639856339,],\n",
+      "[0.4454692602157593,],\n",
+      "[-0.5158112645149231,],\n",
+      "[0.9124316573143005,],\n",
+      "[0.8074878454208374,],\n",
+      "[-0.9386066794395447,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5608106851577759,],\n",
+      "[-0.14065946638584137,],\n",
+      "[-0.6299914121627808,],\n",
+      "[0.1217515841126442,],\n",
+      "[-0.17263594269752502,],\n",
+      "[-0.03727080672979355,],\n",
+      "[-0.3119412064552307,],\n",
+      "[-0.19151651859283447,],\n",
+      "[-0.046674907207489014,],\n",
+      "[-0.0019899490289390087,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3938531279563904,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.676727056503296,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6626602411270142,],\n",
+      "[-0.13359390199184418,],\n",
+      "[0.5313925743103027,],\n",
+      "[-0.9972014427185059,],\n",
+      "[-0.09742648899555206,],\n",
+      "[0.4455065429210663,],\n",
+      "[-0.5154992938041687,],\n",
+      "[0.9126231670379639,],\n",
+      "[0.8075345158576965,],\n",
+      "[-0.9386047124862671,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.39117640256881714,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.8513489961624146,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6626602411270142,],\n",
+      "[-0.13359390199184418,],\n",
+      "[0.5313925743103027,],\n",
+      "[-0.9972014427185059,],\n",
+      "[-0.09742648899555206,],\n",
+      "[0.4455065429210663,],\n",
+      "[-0.5154992938041687,],\n",
+      "[0.9126231670379639,],\n",
+      "[0.8075345158576965,],\n",
+      "[-0.9386047124862671,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.6134431958198547,],\n",
+      "[-0.14373497664928436,],\n",
+      "[-0.6000543236732483,],\n",
+      "[0.11184459179639816,],\n",
+      "[-0.16748011112213135,],\n",
+      "[-0.03491705656051636,],\n",
+      "[-0.2927502989768982,],\n",
+      "[-0.21085302531719208,],\n",
+      "[-0.03825075179338455,],\n",
+      "[-0.0019569306168705225,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.39117640256881714,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.7212860584259033,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6620467901229858,],\n",
+      "[-0.13345016539096832,],\n",
+      "[0.5319926142692566,],\n",
+      "[-0.9973132610321045,],\n",
+      "[-0.09725900739431381,],\n",
+      "[0.44554147124290466,],\n",
+      "[-0.515206515789032,],\n",
+      "[0.9128340482711792,],\n",
+      "[0.8075727820396423,],\n",
+      "[-0.9386027455329895,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3884551227092743,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.537475824356079,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6620467901229858,],\n",
+      "[-0.13345016539096832,],\n",
+      "[0.5319926142692566,],\n",
+      "[-0.9973132610321045,],\n",
+      "[-0.09725900739431381,],\n",
+      "[0.44554147124290466,],\n",
+      "[-0.515206515789032,],\n",
+      "[0.9128340482711792,],\n",
+      "[0.8075727820396423,],\n",
+      "[-0.9386027455329895,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.740168571472168,],\n",
+      "[-0.1629035323858261,],\n",
+      "[-0.6449097990989685,],\n",
+      "[0.1167297437787056,],\n",
+      "[-0.10903051495552063,],\n",
+      "[-0.02253592759370804,],\n",
+      "[-0.1900072991847992,],\n",
+      "[-0.2034318745136261,],\n",
+      "[-0.021931782364845276,],\n",
+      "[-0.001750173163600266,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3884551227092743,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.185892105102539,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.661306619644165,],\n",
+      "[-0.13328726589679718,],\n",
+      "[0.5326375365257263,],\n",
+      "[-0.9974299669265747,],\n",
+      "[-0.09714997559785843,],\n",
+      "[0.44556400179862976,],\n",
+      "[-0.515016496181488,],\n",
+      "[0.9130374789237976,],\n",
+      "[0.8075947165489197,],\n",
+      "[-0.938601016998291,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3852692246437073,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.3028957843780518,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.661306619644165,],\n",
+      "[-0.13328726589679718,],\n",
+      "[0.5326375365257263,],\n",
+      "[-0.9974299669265747,],\n",
+      "[-0.09714997559785843,],\n",
+      "[0.44556400179862976,],\n",
+      "[-0.515016496181488,],\n",
+      "[0.9130374789237976,],\n",
+      "[0.8075947165489197,],\n",
+      "[-0.938601016998291,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.8321822881698608,],\n",
+      "[-0.17373912036418915,],\n",
+      "[-0.6630021929740906,],\n",
+      "[0.1172514259815216,],\n",
+      "[-0.05387640744447708,],\n",
+      "[-0.010788949206471443,],\n",
+      "[-0.09002402424812317,],\n",
+      "[-0.20487558841705322,],\n",
+      "[-0.010124783962965012,],\n",
+      "[-0.0016042840434238315,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3852692246437073,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.6347692012786865,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.660474419593811,],\n",
+      "[-0.1331135332584381,],\n",
+      "[0.533300518989563,],\n",
+      "[-0.9975472092628479,],\n",
+      "[-0.09709610044956207,],\n",
+      "[0.4455747902393341,],\n",
+      "[-0.5149264931678772,],\n",
+      "[0.9132423400878906,],\n",
+      "[0.8076048493385315,],\n",
+      "[-0.9385994076728821,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3816344439983368,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.323455572128296,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.660474419593811,],\n",
+      "[-0.1331135332584381,],\n",
+      "[0.533300518989563,],\n",
+      "[-0.9975472092628479,],\n",
+      "[-0.09709610044956207,],\n",
+      "[0.4455747902393341,],\n",
+      "[-0.5149264931678772,],\n",
+      "[0.9132423400878906,],\n",
+      "[0.8076048493385315,],\n",
+      "[-0.9385994076728821,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.7867519855499268,],\n",
+      "[-0.157746821641922,],\n",
+      "[-0.5891675353050232,],\n",
+      "[0.10241935402154922,],\n",
+      "[-0.05529075115919113,],\n",
+      "[-0.00971552450209856,],\n",
+      "[-0.0806700736284256,],\n",
+      "[-0.233682781457901,],\n",
+      "[-0.014480477198958397,],\n",
+      "[-0.001543051446788013,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3816344439983368,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.6460654735565186,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6596876382827759,],\n",
+      "[-0.13295578956604004,],\n",
+      "[0.5338897109031677,],\n",
+      "[-0.997649610042572,],\n",
+      "[-0.0970408096909523,],\n",
+      "[0.4455845057964325,],\n",
+      "[-0.5148458480834961,],\n",
+      "[0.9134760499000549,],\n",
+      "[0.8076193332672119,],\n",
+      "[-0.9385978579521179,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.37798836827278137,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.3586111068725586,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6596876382827759,],\n",
+      "[-0.13295578956604004,],\n",
+      "[0.5338897109031677,],\n",
+      "[-0.997649610042572,],\n",
+      "[-0.0970408096909523,],\n",
+      "[0.4455845057964325,],\n",
+      "[-0.5148458480834961,],\n",
+      "[0.9134760499000549,],\n",
+      "[0.8076193332672119,],\n",
+      "[-0.9385978579521179,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5987049341201782,],\n",
+      "[-0.11671721935272217,],\n",
+      "[-0.43290573358535767,],\n",
+      "[0.07443776726722717,],\n",
+      "[-0.13016632199287415,],\n",
+      "[-0.022342197597026825,],\n",
+      "[-0.19457465410232544,],\n",
+      "[-0.28364914655685425,],\n",
+      "[-0.0349983274936676,],\n",
+      "[-0.0015205873642116785,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.37798836827278137,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.0715525150299072,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6590889096260071,],\n",
+      "[-0.13283906877040863,],\n",
+      "[0.5343226194381714,],\n",
+      "[-0.9977240562438965,],\n",
+      "[-0.09691064059734344,],\n",
+      "[0.44560685753822327,],\n",
+      "[-0.5146512985229492,],\n",
+      "[0.913759708404541,],\n",
+      "[0.8076543211936951,],\n",
+      "[-0.9385963082313538,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.37491682171821594,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1342180967330933,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6590889096260071,],\n",
+      "[-0.13283906877040863,],\n",
+      "[0.5343226194381714,],\n",
+      "[-0.9977240562438965,],\n",
+      "[-0.09691064059734344,],\n",
+      "[0.44560685753822327,],\n",
+      "[-0.5146512985229492,],\n",
+      "[0.913759708404541,],\n",
+      "[0.8076543211936951,],\n",
+      "[-0.9385963082313538,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.36089423298835754,],\n",
+      "[-0.06910977512598038,],\n",
+      "[-0.2582223117351532,],\n",
+      "[0.04419274255633354,],\n",
+      "[-0.24908572435379028,],\n",
+      "[-0.04313754290342331,],\n",
+      "[-0.38695263862609863,],\n",
+      "[-0.3348022699356079,],\n",
+      "[-0.0607636384665966,],\n",
+      "[-0.0015118721639737487,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.37491682171821594,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1299943923950195,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6587280035018921,],\n",
+      "[-0.13276995718479156,],\n",
+      "[0.5345808267593384,],\n",
+      "[-0.997768223285675,],\n",
+      "[-0.09666155278682709,],\n",
+      "[0.44564998149871826,],\n",
+      "[-0.5142643451690674,],\n",
+      "[0.9140945076942444,],\n",
+      "[0.8077150583267212,],\n",
+      "[-0.9385948181152344,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.37278681993484497,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4075283110141754,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6587280035018921,],\n",
+      "[-0.13276995718479156,],\n",
+      "[0.5345808267593384,],\n",
+      "[-0.997768223285675,],\n",
+      "[-0.09666155278682709,],\n",
+      "[0.44564998149871826,],\n",
+      "[-0.5142643451690674,],\n",
+      "[0.9140945076942444,],\n",
+      "[0.8077150583267212,],\n",
+      "[-0.9385948181152344,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.18191038072109222,],\n",
+      "[-0.03447641059756279,],\n",
+      "[-0.13155058026313782,],\n",
+      "[0.02253354899585247,],\n",
+      "[-0.3550148010253906,],\n",
+      "[-0.061752453446388245,],\n",
+      "[-0.5626010298728943,],\n",
+      "[-0.36973434686660767,],\n",
+      "[-0.0798565000295639,],\n",
+      "[-0.001507492852397263,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.37278681993484497,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.276759147644043,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6585460901260376,],\n",
+      "[-0.132735475897789,],\n",
+      "[0.5347123742103577,],\n",
+      "[-0.9977907538414001,],\n",
+      "[-0.09630654007196426,],\n",
+      "[0.44571173191070557,],\n",
+      "[-0.5137017369270325,],\n",
+      "[0.9144642353057861,],\n",
+      "[0.8077949285507202,],\n",
+      "[-0.938593327999115,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3715100586414337,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.21556539833545685,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6585460901260376,],\n",
+      "[-0.132735475897789,],\n",
+      "[0.5347123742103577,],\n",
+      "[-0.9977907538414001,],\n",
+      "[-0.09630654007196426,],\n",
+      "[0.44571173191070557,],\n",
+      "[-0.5137017369270325,],\n",
+      "[0.9144642353057861,],\n",
+      "[0.8077949285507202,],\n",
+      "[-0.938593327999115,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.10822043567895889,],\n",
+      "[-0.02042161114513874,],\n",
+      "[-0.08050507307052612,],\n",
+      "[0.01386394165456295,],\n",
+      "[-0.3982177972793579,],\n",
+      "[-0.06907863914966583,],\n",
+      "[-0.6345381140708923,],\n",
+      "[-0.3830758333206177,],\n",
+      "[-0.08762481808662415,],\n",
+      "[-0.0014993607765063643,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3715100586414337,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9285807013511658,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6584378480911255,],\n",
+      "[-0.13271506130695343,],\n",
+      "[0.5347929000854492,],\n",
+      "[-0.9978046417236328,],\n",
+      "[-0.09590832144021988,],\n",
+      "[0.44578081369400024,],\n",
+      "[-0.5130671858787537,],\n",
+      "[0.9148473143577576,],\n",
+      "[0.80788254737854,],\n",
+      "[-0.9385918378829956,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3705814778804779,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3664473295211792,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6584378480911255,],\n",
+      "[-0.13271506130695343,],\n",
+      "[0.5347929000854492,],\n",
+      "[-0.9978046417236328,],\n",
+      "[-0.09590832144021988,],\n",
+      "[0.44578081369400024,],\n",
+      "[-0.5130671858787537,],\n",
+      "[0.9148473143577576,],\n",
+      "[0.80788254737854,],\n",
+      "[-0.9385918378829956,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.11280979216098785,],\n",
+      "[-0.021332671865820885,],\n",
+      "[-0.08759402483701706,],\n",
+      "[0.015216326341032982,],\n",
+      "[-0.36598336696624756,],\n",
+      "[-0.06239846348762512,],\n",
+      "[-0.5775666236877441,],\n",
+      "[-0.3809010684490204,],\n",
+      "[-0.08728288114070892,],\n",
+      "[-0.0014846906997263432,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3705814778804779,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.2106975317001343,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6583250164985657,],\n",
+      "[-0.13269372284412384,],\n",
+      "[0.534880518913269,],\n",
+      "[-0.9978198409080505,],\n",
+      "[-0.09554234147071838,],\n",
+      "[0.4458432197570801,],\n",
+      "[-0.5124896168708801,],\n",
+      "[0.9152281880378723,],\n",
+      "[0.8079698085784912,],\n",
+      "[-0.9385903477668762,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.36937078833580017,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8836364150047302,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6583250164985657,],\n",
+      "[-0.13269372284412384,],\n",
+      "[0.534880518913269,],\n",
+      "[-0.9978198409080505,],\n",
+      "[-0.09554234147071838,],\n",
+      "[0.4458432197570801,],\n",
+      "[-0.5124896168708801,],\n",
+      "[0.9152281880378723,],\n",
+      "[0.8079698085784912,],\n",
+      "[-0.9385903477668762,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.13742807507514954,],\n",
+      "[-0.02625006064772606,],\n",
+      "[-0.11273419111967087,],\n",
+      "[0.019802967086434364,],\n",
+      "[-0.28969794511795044,],\n",
+      "[-0.04714852571487427,],\n",
+      "[-0.4435634911060333,],\n",
+      "[-0.3739319443702698,],\n",
+      "[-0.08500197529792786,],\n",
+      "[-0.0014675717102363706,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.36937078833580017,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8800380229949951,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6581875681877136,],\n",
+      "[-0.13266746699810028,],\n",
+      "[0.5349932312965393,],\n",
+      "[-0.997839629650116,],\n",
+      "[-0.0952526405453682,],\n",
+      "[0.4458903670310974,],\n",
+      "[-0.5120460391044617,],\n",
+      "[0.9156021475791931,],\n",
+      "[0.8080548048019409,],\n",
+      "[-0.9385888576507568,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3674907386302948,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.526879906654358,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6581875681877136,],\n",
+      "[-0.13266746699810028,],\n",
+      "[0.5349932312965393,],\n",
+      "[-0.997839629650116,],\n",
+      "[-0.0952526405453682,],\n",
+      "[0.4458903670310974,],\n",
+      "[-0.5120460391044617,],\n",
+      "[0.9156021475791931,],\n",
+      "[0.8080548048019409,],\n",
+      "[-0.9385888576507568,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.13954786956310272,],\n",
+      "[-0.027153506875038147,],\n",
+      "[-0.12211408466100693,],\n",
+      "[0.02173241414129734,],\n",
+      "[-0.22511614859104156,],\n",
+      "[-0.03380211815237999,],\n",
+      "[-0.3286017179489136,],\n",
+      "[-0.3704655170440674,],\n",
+      "[-0.08498015254735947,],\n",
+      "[-0.0014533979119732976,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3674907386302948,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.47133731842041,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6580480337142944,],\n",
+      "[-0.1326403170824051,],\n",
+      "[0.5351153612136841,],\n",
+      "[-0.997861385345459,],\n",
+      "[-0.09502752125263214,],\n",
+      "[0.44592416286468506,],\n",
+      "[-0.5117174386978149,],\n",
+      "[0.9159725904464722,],\n",
+      "[0.8081398010253906,],\n",
+      "[-0.9385874271392822,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.36501941084861755,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.6613308191299438,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6580480337142944,],\n",
+      "[-0.1326403170824051,],\n",
+      "[0.5351153612136841,],\n",
+      "[-0.997861385345459,],\n",
+      "[-0.09502752125263214,],\n",
+      "[0.44592416286468506,],\n",
+      "[-0.5117174386978149,],\n",
+      "[0.9159725904464722,],\n",
+      "[0.8081398010253906,],\n",
+      "[-0.9385874271392822,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.11109857261180878,],\n",
+      "[-0.02217385731637478,],\n",
+      "[-0.10445858538150787,],\n",
+      "[0.018864868208765984,],\n",
+      "[-0.2179175615310669,],\n",
+      "[-0.03125779330730438,],\n",
+      "[-0.31295180320739746,],\n",
+      "[-0.3730130195617676,],\n",
+      "[-0.08782752603292465,],\n",
+      "[-0.0014446871355175972,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.36501941084861755,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.577852964401245,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6579369306564331,],\n",
+      "[-0.13261814415454865,],\n",
+      "[0.5352198481559753,],\n",
+      "[-0.997880220413208,],\n",
+      "[-0.09480960667133331,],\n",
+      "[0.44595542550086975,],\n",
+      "[-0.5114045143127441,],\n",
+      "[0.9163455963134766,],\n",
+      "[0.8082276582717896,],\n",
+      "[-0.9385859966278076,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3624415695667267,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.0988463163375854,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6579369306564331,],\n",
+      "[-0.13261814415454865,],\n",
+      "[0.5352198481559753,],\n",
+      "[-0.997880220413208,],\n",
+      "[-0.09480960667133331,],\n",
+      "[0.44595542550086975,],\n",
+      "[-0.5114045143127441,],\n",
+      "[0.9163455963134766,],\n",
+      "[0.8082276582717896,],\n",
+      "[-0.9385859966278076,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.06836346536874771,],\n",
+      "[-0.014042040333151817,],\n",
+      "[-0.06939537078142166,],\n",
+      "[0.012735673226416111,],\n",
+      "[-0.2762259244918823,],\n",
+      "[-0.041290126740932465,],\n",
+      "[-0.412789523601532,],\n",
+      "[-0.37931257486343384,],\n",
+      "[-0.09187787026166916,],\n",
+      "[-0.0014408595161512494,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3624415695667267,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.0965187549591064,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6578685641288757,],\n",
+      "[-0.13260410726070404,],\n",
+      "[0.5352892279624939,],\n",
+      "[-0.9978929758071899,],\n",
+      "[-0.09453338384628296,],\n",
+      "[0.4459967017173767,],\n",
+      "[-0.5109917521476746,],\n",
+      "[0.9167249202728271,],\n",
+      "[0.8083195090293884,],\n",
+      "[-0.938584566116333,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3603450655937195,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4152942895889282,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6578685641288757,],\n",
+      "[-0.13260410726070404,],\n",
+      "[0.5352892279624939,],\n",
+      "[-0.9978929758071899,],\n",
+      "[-0.09453338384628296,],\n",
+      "[0.4459967017173767,],\n",
+      "[-0.5109917521476746,],\n",
+      "[0.9167249202728271,],\n",
+      "[0.8083195090293884,],\n",
+      "[-0.938584566116333,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0316094234585762,],\n",
+      "[-0.006679167039692402,],\n",
+      "[-0.03473832085728645,],\n",
+      "[0.006486787926405668,],\n",
+      "[-0.3674730658531189,],\n",
+      "[-0.057907260954380035,],\n",
+      "[-0.5727624893188477,],\n",
+      "[-0.385593980550766,],\n",
+      "[-0.09524527192115784,],\n",
+      "[-0.0014398780185729265,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3603450655937195,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.2888667583465576,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6578369736671448,],\n",
+      "[-0.1325974315404892,],\n",
+      "[0.535323977470398,],\n",
+      "[-0.9978994727134705,],\n",
+      "[-0.09416591376066208,],\n",
+      "[0.446054607629776,],\n",
+      "[-0.5104190111160278,],\n",
+      "[0.9171105027198792,],\n",
+      "[0.8084147572517395,],\n",
+      "[-0.9385831356048584,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3590562045574188,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09632387012243271,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6578369736671448,],\n",
+      "[-0.1325974315404892,],\n",
+      "[0.535323977470398,],\n",
+      "[-0.9978994727134705,],\n",
+      "[-0.09416591376066208,],\n",
+      "[0.446054607629776,],\n",
+      "[-0.5104190111160278,],\n",
+      "[0.9171105027198792,],\n",
+      "[0.8084147572517395,],\n",
+      "[-0.9385831356048584,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.011402584612369537,],\n",
+      "[-0.002472421620041132,],\n",
+      "[-0.013587803579866886,],\n",
+      "[0.002584459725767374,],\n",
+      "[-0.4402979016304016,],\n",
+      "[-0.0714193657040596,],\n",
+      "[-0.701919674873352,],\n",
+      "[-0.3892700672149658,],\n",
+      "[-0.0970456674695015,],\n",
+      "[-0.0014397903578355908,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3590562045574188,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6207221746444702,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6578255891799927,],\n",
+      "[-0.13259495794773102,],\n",
+      "[0.5353375673294067,],\n",
+      "[-0.9979020357131958,],\n",
+      "[-0.09372561424970627,],\n",
+      "[0.4461260139942169,],\n",
+      "[-0.5097171068191528,],\n",
+      "[0.9174997806549072,],\n",
+      "[0.8085117936134338,],\n",
+      "[-0.9385817050933838,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3584354817867279,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.057123493403196335,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6578255891799927,],\n",
+      "[-0.13259495794773102,],\n",
+      "[0.5353375673294067,],\n",
+      "[-0.9979020357131958,],\n",
+      "[-0.09372561424970627,],\n",
+      "[0.4461260139942169,],\n",
+      "[-0.5097171068191528,],\n",
+      "[0.9174997806549072,],\n",
+      "[0.8085117936134338,],\n",
+      "[-0.9385817050933838,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0065568359568715096,],\n",
+      "[-0.0014557581162080169,],\n",
+      "[-0.008480624295771122,],\n",
+      "[0.0016444330103695393,],\n",
+      "[-0.4562321901321411,],\n",
+      "[-0.07420119643211365,],\n",
+      "[-0.7302073836326599,],\n",
+      "[-0.38984644412994385,],\n",
+      "[-0.09747195988893509,],\n",
+      "[-0.0014395163161680102,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3584354817867279,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4780106246471405,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6578190326690674,],\n",
+      "[-0.13259349763393402,],\n",
+      "[0.5353460311889648,],\n",
+      "[-0.9979037046432495,],\n",
+      "[-0.09326938539743423,],\n",
+      "[0.4462002217769623,],\n",
+      "[-0.5089868903160095,],\n",
+      "[0.9178896546363831,],\n",
+      "[0.8086092472076416,],\n",
+      "[-0.9385802745819092,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.35795748233795166,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.21680709719657898,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6578190326690674,],\n",
+      "[-0.13259349763393402,],\n",
+      "[0.5353460311889648,],\n",
+      "[-0.9979037046432495,],\n",
+      "[-0.09326938539743423,],\n",
+      "[0.4462002217769623,],\n",
+      "[-0.5089868903160095,],\n",
+      "[0.9178896546363831,],\n",
+      "[0.8086092472076416,],\n",
+      "[-0.9385802745819092,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.009513515047729015,],\n",
+      "[-0.0021628723479807377,],\n",
+      "[-0.013368302956223488,],\n",
+      "[0.0026439528446644545,],\n",
+      "[-0.4119870662689209,],\n",
+      "[-0.06507056951522827,],\n",
+      "[-0.6507904529571533,],\n",
+      "[-0.38831833004951477,],\n",
+      "[-0.09723374247550964,],\n",
+      "[-0.0014388490235432982,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.35795748233795166,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9312511682510376,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6578094959259033,],\n",
+      "[-0.1325913369655609,],\n",
+      "[0.5353593826293945,],\n",
+      "[-0.9979063272476196,],\n",
+      "[-0.09285739809274673,],\n",
+      "[0.4462652802467346,],\n",
+      "[-0.5083361268043518,],\n",
+      "[0.9182779788970947,],\n",
+      "[0.8087064623832703,],\n",
+      "[-0.9385788440704346,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.35702621936798096,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7155236005783081,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6578094959259033,],\n",
+      "[-0.1325913369655609,],\n",
+      "[0.5353593826293945,],\n",
+      "[-0.9979063272476196,],\n",
+      "[-0.09285739809274673,],\n",
+      "[0.4462652802467346,],\n",
+      "[-0.5083361268043518,],\n",
+      "[0.9182779788970947,],\n",
+      "[0.8087064623832703,],\n",
+      "[-0.9385788440704346,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.012842817232012749,],\n",
+      "[-0.0029974388889968395,],\n",
+      "[-0.019626695662736893,],\n",
+      "[0.003959818743169308,],\n",
+      "[-0.3402442932128906,],\n",
+      "[-0.050045765936374664,],\n",
+      "[-0.5213673710823059,],\n",
+      "[-0.3860744535923004,],\n",
+      "[-0.09697528183460236,],\n",
+      "[-0.001438043313100934,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.35702621936798096,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.6917715072631836,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577966809272766,],\n",
+      "[-0.13258834183216095,],\n",
+      "[0.5353789925575256,],\n",
+      "[-0.9979102611541748,],\n",
+      "[-0.09251715242862701,],\n",
+      "[0.44631531834602356,],\n",
+      "[-0.5078147649765015,],\n",
+      "[0.9186640381813049,],\n",
+      "[0.8088034391403198,],\n",
+      "[-0.93857741355896,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.35533446073532104,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.3067326545715332,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577966809272766,],\n",
+      "[-0.13258834183216095,],\n",
+      "[0.5353789925575256,],\n",
+      "[-0.9979102611541748,],\n",
+      "[-0.09251715242862701,],\n",
+      "[0.44631531834602356,],\n",
+      "[-0.5078147649765015,],\n",
+      "[0.9186640381813049,],\n",
+      "[0.8088034391403198,],\n",
+      "[-0.93857741355896,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.01287386380136013,],\n",
+      "[-0.0030960016883909702,],\n",
+      "[-0.02141774445772171,],\n",
+      "[0.004407576750963926,],\n",
+      "[-0.2887730002403259,],\n",
+      "[-0.038646548986434937,],\n",
+      "[-0.4257259964942932,],\n",
+      "[-0.38408076763153076,],\n",
+      "[-0.09698621183633804,],\n",
+      "[-0.001437382074072957,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.35533446073532104,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.2862460613250732,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577838063240051,],\n",
+      "[-0.13258524239063263,],\n",
+      "[0.535400390625,],\n",
+      "[-0.9979146718978882,],\n",
+      "[-0.09222838282585144,],\n",
+      "[0.4463539719581604,],\n",
+      "[-0.5073890686035156,],\n",
+      "[0.9190481305122375,],\n",
+      "[0.8089004158973694,],\n",
+      "[-0.9385759830474854,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3530482053756714,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.3730260133743286,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577838063240051,],\n",
+      "[-0.13258524239063263,],\n",
+      "[0.535400390625,],\n",
+      "[-0.9979146718978882,],\n",
+      "[-0.09222838282585144,],\n",
+      "[0.4463539719581604,],\n",
+      "[-0.5073890686035156,],\n",
+      "[0.9190481305122375,],\n",
+      "[0.8089004158973694,],\n",
+      "[-0.9385759830474854,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.009776473045349121,],\n",
+      "[-0.002429170999675989,],\n",
+      "[-0.017717687413096428,],\n",
+      "[0.003718509105965495,],\n",
+      "[-0.29043006896972656,],\n",
+      "[-0.03781958296895027,],\n",
+      "[-0.42308419942855835,],\n",
+      "[-0.38266468048095703,],\n",
+      "[-0.09723877906799316,],\n",
+      "[-0.0014369847485795617,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3530482053756714,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.3435254096984863,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.657774031162262,],\n",
+      "[-0.13258281350135803,],\n",
+      "[0.5354180932044983,],\n",
+      "[-0.9979183673858643,],\n",
+      "[-0.09193795174360275,],\n",
+      "[0.4463917911052704,],\n",
+      "[-0.5069659948348999,],\n",
+      "[0.9194307923316956,],\n",
+      "[0.808997631072998,],\n",
+      "[-0.9385745525360107,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3507046699523926,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8276777863502502,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.657774031162262,],\n",
+      "[-0.13258281350135803,],\n",
+      "[0.5354180932044983,],\n",
+      "[-0.9979183673858643,],\n",
+      "[-0.09193795174360275,],\n",
+      "[0.4463917911052704,],\n",
+      "[-0.5069659948348999,],\n",
+      "[0.9194307923316956,],\n",
+      "[0.808997631072998,],\n",
+      "[-0.9385745525360107,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0056192586198449135,],\n",
+      "[-0.0014429338043555617,],\n",
+      "[-0.01109496969729662,],\n",
+      "[0.002374940551817417,],\n",
+      "[-0.3437070846557617,],\n",
+      "[-0.04772019013762474,],\n",
+      "[-0.5153163075447083,],\n",
+      "[-0.38176435232162476,],\n",
+      "[-0.09756259620189667,],\n",
+      "[-0.0014368268894031644,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3507046699523926,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8195382356643677,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577684283256531,],\n",
+      "[-0.13258136808872223,],\n",
+      "[0.5354291796684265,],\n",
+      "[-0.9979207515716553,],\n",
+      "[-0.09159424155950546,],\n",
+      "[0.4464395046234131,],\n",
+      "[-0.5064506530761719,],\n",
+      "[0.9198125600814819,],\n",
+      "[0.8090952038764954,],\n",
+      "[-0.9385731220245361,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.34888511896133423,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.26204097270965576,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577684283256531,],\n",
+      "[-0.13258136808872223,],\n",
+      "[0.5354291796684265,],\n",
+      "[-0.9979207515716553,],\n",
+      "[-0.09159424155950546,],\n",
+      "[0.4464395046234131,],\n",
+      "[-0.5064506530761719,],\n",
+      "[0.9198125600814819,],\n",
+      "[0.8090952038764954,],\n",
+      "[-0.9385731220245361,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0023397128097712994,],\n",
+      "[-0.0006195122259669006,],\n",
+      "[-0.005031326785683632,],\n",
+      "[0.0010988559806719422,],\n",
+      "[-0.41640037298202515,],\n",
+      "[-0.06211128458380699,],\n",
+      "[-0.6478341817855835,],\n",
+      "[-0.3812233805656433,],\n",
+      "[-0.09781008958816528,],\n",
+      "[-0.0014368020929396152,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.34888511896133423,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0237987041473389,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577661037445068,],\n",
+      "[-0.1325807422399521,],\n",
+      "[0.5354341864585876,],\n",
+      "[-0.9979218244552612,],\n",
+      "[-0.09117784351110458,],\n",
+      "[0.44650161266326904,],\n",
+      "[-0.5058028101921082,],\n",
+      "[0.9201937913894653,],\n",
+      "[0.8091930150985718,],\n",
+      "[-0.9385716915130615,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.34786131978034973,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.046505436301231384,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577661037445068,],\n",
+      "[-0.1325807422399521,],\n",
+      "[0.5354341864585876,],\n",
+      "[-0.9979218244552612,],\n",
+      "[-0.09117784351110458,],\n",
+      "[0.44650161266326904,],\n",
+      "[-0.5058028101921082,],\n",
+      "[0.9201937913894653,],\n",
+      "[0.8091930150985718,],\n",
+      "[-0.9385716915130615,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0007292124209925532,],\n",
+      "[-0.00019843610061798245,],\n",
+      "[-0.0017068340675905347,],\n",
+      "[0.00038056436460465193,],\n",
+      "[-0.46749550104141235,],\n",
+      "[-0.07250605523586273,],\n",
+      "[-0.7434905171394348,],\n",
+      "[-0.38088175654411316,],\n",
+      "[-0.09792827814817429,],\n",
+      "[-0.0014367958065122366,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.34786131978034973,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4313022196292877,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577653884887695,],\n",
+      "[-0.13258054852485657,],\n",
+      "[0.5354359149932861,],\n",
+      "[-0.9979221820831299,],\n",
+      "[-0.09071034938097,],\n",
+      "[0.4465741217136383,],\n",
+      "[-0.5050593018531799,],\n",
+      "[0.9205746650695801,],\n",
+      "[0.8092909455299377,],\n",
+      "[-0.9385702610015869,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.347430020570755,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03881211206316948,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577653884887695,],\n",
+      "[-0.13258054852485657,],\n",
+      "[0.5354359149932861,],\n",
+      "[-0.9979221820831299,],\n",
+      "[-0.09071034938097,],\n",
+      "[0.4465741217136383,],\n",
+      "[-0.5050593018531799,],\n",
+      "[0.9205746650695801,],\n",
+      "[0.8092909455299377,],\n",
+      "[-0.9385702610015869,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0004927366389892995,],\n",
+      "[-0.00013747044431511313,],\n",
+      "[-0.00125477509573102,],\n",
+      "[0.00028574257157742977,],\n",
+      "[-0.4715503454208374,],\n",
+      "[-0.07316210865974426,],\n",
+      "[-0.7508536577224731,],\n",
+      "[-0.3805437386035919,],\n",
+      "[-0.09794562309980392,],\n",
+      "[-0.0014367903349921107,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.347430020570755,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3940157890319824,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577649116516113,],\n",
+      "[-0.13258041441440582,],\n",
+      "[0.5354371666908264,],\n",
+      "[-0.9979224801063538,],\n",
+      "[-0.09023880213499069,],\n",
+      "[0.4466472864151001,],\n",
+      "[-0.5043084621429443,],\n",
+      "[0.9209551811218262,],\n",
+      "[0.8093888759613037,],\n",
+      "[-0.9385688304901123,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3470360040664673,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.21489785611629486,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577649116516113,],\n",
+      "[-0.13258041441440582,],\n",
+      "[0.5354371666908264,],\n",
+      "[-0.9979224801063538,],\n",
+      "[-0.09023880213499069,],\n",
+      "[0.4466472864151001,],\n",
+      "[-0.5043084621429443,],\n",
+      "[0.9209551811218262,],\n",
+      "[0.8093888759613037,],\n",
+      "[-0.9385688304901123,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0008573830127716064,],\n",
+      "[-0.00024520998704247177,],\n",
+      "[-0.002375254174694419,],\n",
+      "[0.000552481331396848,],\n",
+      "[-0.4321736991405487,],\n",
+      "[-0.06429485976696014,],\n",
+      "[-0.6748409867286682,],\n",
+      "[-0.3799358010292053,],\n",
+      "[-0.09792068600654602,],\n",
+      "[-0.001436749123968184,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3470360040664673,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9271427989006042,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577640771865845,],\n",
+      "[-0.13258017599582672,],\n",
+      "[0.5354395508766174,],\n",
+      "[-0.9979230165481567,],\n",
+      "[-0.08980663120746613,],\n",
+      "[0.44671157002449036,],\n",
+      "[-0.5036336183547974,],\n",
+      "[0.9213351011276245,],\n",
+      "[0.8094868063926697,],\n",
+      "[-0.9385673999786377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3461088538169861,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7223889827728271,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577640771865845,],\n",
+      "[-0.13258017599582672,],\n",
+      "[0.5354395508766174,],\n",
+      "[-0.9979230165481567,],\n",
+      "[-0.08980663120746613,],\n",
+      "[0.44671157002449036,],\n",
+      "[-0.5036336183547974,],\n",
+      "[0.9213351011276245,],\n",
+      "[0.8094868063926697,],\n",
+      "[-0.9385673999786377,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0011621778830885887,],\n",
+      "[-0.00034146735561080277,],\n",
+      "[-0.003503896528854966,],\n",
+      "[0.0008321304921992123,],\n",
+      "[-0.3785366714000702,],\n",
+      "[-0.05184965580701828,],\n",
+      "[-0.5689355134963989,],\n",
+      "[-0.3787952959537506,],\n",
+      "[-0.09790071099996567,],\n",
+      "[-0.0014367084950208664,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3461088538169861,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.6998677253723145,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577629446983337,],\n",
+      "[-0.13257983326911926,],\n",
+      "[0.5354430675506592,],\n",
+      "[-0.9979238510131836,],\n",
+      "[-0.08942809700965881,],\n",
+      "[0.44676342606544495,],\n",
+      "[-0.5030646920204163,],\n",
+      "[0.9217138886451721,],\n",
+      "[0.8095847368240356,],\n",
+      "[-0.9385659694671631,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3444089889526367,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2500072717666626,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577629446983337,],\n",
+      "[-0.13257983326911926,],\n",
+      "[0.5354430675506592,],\n",
+      "[-0.9979238510131836,],\n",
+      "[-0.08942809700965881,],\n",
+      "[0.44676342606544495,],\n",
+      "[-0.5030646920204163,],\n",
+      "[0.9217138886451721,],\n",
+      "[0.8095847368240356,],\n",
+      "[-0.9385659694671631,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0011300357291474938,],\n",
+      "[-0.000342156330589205,],\n",
+      "[-0.0037098987959325314,],\n",
+      "[0.0008990289061330259,],\n",
+      "[-0.34657877683639526,],\n",
+      "[-0.04373185336589813,],\n",
+      "[-0.4993775188922882,],\n",
+      "[-0.3770703971385956,],\n",
+      "[-0.09790438413619995,],\n",
+      "[-0.0014366836985573173,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3444089889526367,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.236072063446045,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.657761812210083,],\n",
+      "[-0.1325794905424118,],\n",
+      "[0.5354467630386353,],\n",
+      "[-0.9979247450828552,],\n",
+      "[-0.08908151835203171,],\n",
+      "[0.4468071460723877,],\n",
+      "[-0.502565324306488,],\n",
+      "[0.9220909476280212,],\n",
+      "[0.8096826672554016,],\n",
+      "[-0.9385645389556885,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.34217292070388794,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2134400606155396,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.657761812210083,],\n",
+      "[-0.1325794905424118,],\n",
+      "[0.5354467630386353,],\n",
+      "[-0.9979247450828552,],\n",
+      "[-0.08908151835203171,],\n",
+      "[0.4468071460723877,],\n",
+      "[-0.502565324306488,],\n",
+      "[0.9220909476280212,],\n",
+      "[0.8096826672554016,],\n",
+      "[-0.9385645389556885,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0008228975348174572,],\n",
+      "[-0.00025729788467288017,],\n",
+      "[-0.0029429157730191946,],\n",
+      "[0.0007273925002664328,],\n",
+      "[-0.35619139671325684,],\n",
+      "[-0.04479164257645607,],\n",
+      "[-0.5077866911888123,],\n",
+      "[-0.37499117851257324,],\n",
+      "[-0.09792449325323105,],\n",
+      "[-0.001436662394553423,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.34217292070388794,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.2031280994415283,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577609777450562,],\n",
+      "[-0.1325792372226715,],\n",
+      "[0.5354496836662292,],\n",
+      "[-0.9979254603385925,],\n",
+      "[-0.08872532844543457,],\n",
+      "[0.4468519389629364,],\n",
+      "[-0.5020575523376465,],\n",
+      "[0.9224659204483032,],\n",
+      "[0.8097805976867676,],\n",
+      "[-0.9385631084442139,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.33996978402137756,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6510122418403625,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577609777450562,],\n",
+      "[-0.1325792372226715,],\n",
+      "[0.5354496836662292,],\n",
+      "[-0.9979254603385925,],\n",
+      "[-0.08872532844543457,],\n",
+      "[0.4468519389629364,],\n",
+      "[-0.5020575523376465,],\n",
+      "[0.9224659204483032,],\n",
+      "[0.8097805976867676,],\n",
+      "[-0.9385631084442139,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0004454603185877204,],\n",
+      "[-0.0001438007311662659,],\n",
+      "[-0.0017353215953335166,],\n",
+      "[0.00043747498420998454,],\n",
+      "[-0.400298148393631,],\n",
+      "[-0.05403825640678406,],\n",
+      "[-0.5896078944206238,],\n",
+      "[-0.3730219900608063,],\n",
+      "[-0.09794928878545761,],\n",
+      "[-0.0014366634422913194,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.33996978402137756,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.613703966140747,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577605605125427,],\n",
+      "[-0.13257908821105957,],\n",
+      "[0.5354514122009277,],\n",
+      "[-0.997925877571106,],\n",
+      "[-0.08832503110170364,],\n",
+      "[0.4469059705734253,],\n",
+      "[-0.5014679431915283,],\n",
+      "[0.9228389263153076,],\n",
+      "[0.8098785281181335,],\n",
+      "[-0.9385616779327393,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.338356077671051,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1673566997051239,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577605605125427,],\n",
+      "[-0.13257908821105957,],\n",
+      "[0.5354514122009277,],\n",
+      "[-0.997925877571106,],\n",
+      "[-0.08832503110170364,],\n",
+      "[0.4469059705734253,],\n",
+      "[-0.5014679431915283,],\n",
+      "[0.9228389263153076,],\n",
+      "[0.8098785281181335,],\n",
+      "[-0.9385616779327393,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00016692061035428196,],\n",
+      "[-5.549983325181529e-05,],\n",
+      "[-0.0007079351926222444,],\n",
+      "[0.00018211905262432992,],\n",
+      "[-0.4524041414260864,],\n",
+      "[-0.06573793292045593,],\n",
+      "[-0.6965022683143616,],\n",
+      "[-0.37159988284111023,],\n",
+      "[-0.09796614199876785,],\n",
+      "[-0.0014366639079526067,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.338356077671051,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8181865215301514,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577603816986084,],\n",
+      "[-0.1325790286064148,],\n",
+      "[0.535452127456665,],\n",
+      "[-0.9979260563850403,],\n",
+      "[-0.08787262439727783,],\n",
+      "[0.44697171449661255,],\n",
+      "[-0.5007714629173279,],\n",
+      "[0.9232105016708374,],\n",
+      "[0.8099765181541443,],\n",
+      "[-0.9385602474212646,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.33753788471221924,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.021315814927220345,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577603816986084,],\n",
+      "[-0.1325790286064148,],\n",
+      "[0.535452127456665,],\n",
+      "[-0.9979260563850403,],\n",
+      "[-0.08787262439727783,],\n",
+      "[0.44697171449661255,],\n",
+      "[-0.5007714629173279,],\n",
+      "[0.9232105016708374,],\n",
+      "[0.8099765181541443,],\n",
+      "[-0.9385602474212646,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[4.4026088289683685e-05,],\n",
+      "[-1.5030132999527268e-05,],\n",
+      "[-0.00020314462017267942,],\n",
+      "[5.336306639946997e-05,],\n",
+      "[-0.48436209559440613,],\n",
+      "[-0.07315947860479355,],\n",
+      "[-0.765468955039978,],\n",
+      "[-0.37087443470954895,],\n",
+      "[-0.09797348827123642,],\n",
+      "[-0.0014366600662469864,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.33753788471221924,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.2919989228248596,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577603220939636,],\n",
+      "[-0.1325790137052536,],\n",
+      "[0.5354523062705994,],\n",
+      "[-0.9979261159896851,],\n",
+      "[-0.08738826215267181,],\n",
+      "[0.44704487919807434,],\n",
+      "[-0.5000060200691223,],\n",
+      "[0.9235813617706299,],\n",
+      "[0.810074508190155,],\n",
+      "[-0.93855881690979,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3372458815574646,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.030117271468043327,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577603220939636,],\n",
+      "[-0.1325790137052536,],\n",
+      "[0.5354523062705994,],\n",
+      "[-0.9979261159896851,],\n",
+      "[-0.08738826215267181,],\n",
+      "[0.44704487919807434,],\n",
+      "[-0.5000060200691223,],\n",
+      "[0.9235813617706299,],\n",
+      "[0.810074508190155,],\n",
+      "[-0.93855881690979,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.8675210817018524e-05,],\n",
+      "[-1.3529095667763613e-05,],\n",
+      "[-0.000194057502085343,],\n",
+      "[5.207584399613552e-05,],\n",
+      "[-0.4820770025253296,],\n",
+      "[-0.072471484541893,],\n",
+      "[-0.7597804069519043,],\n",
+      "[-0.3705825209617615,],\n",
+      "[-0.09797420352697372,],\n",
+      "[-0.0014366591349244118,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3372458815574646,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.34708672761917114,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577602624893188,],\n",
+      "[-0.1325789988040924,],\n",
+      "[0.5354524850845337,],\n",
+      "[-0.9979261755943298,],\n",
+      "[-0.08690618723630905,],\n",
+      "[0.4471173584461212,],\n",
+      "[-0.4992462396621704,],\n",
+      "[0.9239519238471985,],\n",
+      "[0.8101724982261658,],\n",
+      "[-0.9385573863983154,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3368988037109375,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22218340635299683,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577602624893188,],\n",
+      "[-0.1325789988040924,],\n",
+      "[0.5354524850845337,],\n",
+      "[-0.9979261755943298,],\n",
+      "[-0.08690618723630905,],\n",
+      "[0.4471173584461212,],\n",
+      "[-0.4992462396621704,],\n",
+      "[0.9239519238471985,],\n",
+      "[0.8101724982261658,],\n",
+      "[-0.9385573863983154,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.763127359794453e-05,],\n",
+      "[-2.783213676593732e-05,],\n",
+      "[-0.0004235919041093439,],\n",
+      "[0.00011611827358137816,],\n",
+      "[-0.4521245062351227,],\n",
+      "[-0.06480016559362411,],\n",
+      "[-0.690732479095459,],\n",
+      "[-0.37019678950309753,],\n",
+      "[-0.09797194600105286,],\n",
+      "[-0.001436658320017159,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3368988037109375,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9427262544631958,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577602028846741,],\n",
+      "[-0.13257896900177002,],\n",
+      "[0.5354529023170471,],\n",
+      "[-0.9979262948036194,],\n",
+      "[-0.08645406365394592,],\n",
+      "[0.44718214869499207,],\n",
+      "[-0.4985555112361908,],\n",
+      "[0.9243221282958984,],\n",
+      "[0.8102704882621765,],\n",
+      "[-0.9385559558868408,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.33595606684684753,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7313570380210876,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577602028846741,],\n",
+      "[-0.13257896900177002,],\n",
+      "[0.5354529023170471,],\n",
+      "[-0.9979262948036194,],\n",
+      "[-0.08645406365394592,],\n",
+      "[0.44718214869499207,],\n",
+      "[-0.4985555112361908,],\n",
+      "[0.9243221282958984,],\n",
+      "[0.8102704882621765,],\n",
+      "[-0.9385559558868408,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00010408549860585481,],\n",
+      "[-3.83325204893481e-05,],\n",
+      "[-0.0006179207703098655,],\n",
+      "[0.00017294184362981468,],\n",
+      "[-0.41684168577194214,],\n",
+      "[-0.055380679666996,],\n",
+      "[-0.6050458550453186,],\n",
+      "[-0.3692214787006378,],\n",
+      "[-0.09797041118144989,],\n",
+      "[-0.001436656923033297,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.33595606684684753,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7103902101516724,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577600836753845,],\n",
+      "[-0.13257892429828644,],\n",
+      "[0.5354534983634949,],\n",
+      "[-0.9979264736175537,],\n",
+      "[-0.08603721857070923,],\n",
+      "[0.4472375214099884,],\n",
+      "[-0.4979504644870758,],\n",
+      "[0.924691379070282,],\n",
+      "[0.8103684782981873,],\n",
+      "[-0.9385545253753662,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3342456817626953,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1850718259811401,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577600836753845,],\n",
+      "[-0.13257892429828644,],\n",
+      "[0.5354534983634949,],\n",
+      "[-0.9979264736175537,],\n",
+      "[-0.08603721857070923,],\n",
+      "[0.4472375214099884,],\n",
+      "[-0.4979504644870758,],\n",
+      "[0.924691379070282,],\n",
+      "[0.8103684782981873,],\n",
+      "[-0.9385545253753662,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[9.791182674234733e-05,],\n",
+      "[-3.7148860428715125e-05,],\n",
+      "[-0.0006328417221084237,],\n",
+      "[0.0001807081134757027,],\n",
+      "[-0.39987772703170776,],\n",
+      "[-0.05020710825920105,],\n",
+      "[-0.5544456839561462,],\n",
+      "[-0.3675113916397095,],\n",
+      "[-0.09797109663486481,],\n",
+      "[-0.0014366551768034697,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3342456817626953,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1772172451019287,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.657759964466095,],\n",
+      "[-0.13257889449596405,],\n",
+      "[0.5354541540145874,],\n",
+      "[-0.997926652431488,],\n",
+      "[-0.08563733845949173,],\n",
+      "[0.44728773832321167,],\n",
+      "[-0.4973960220813751,],\n",
+      "[0.925058901309967,],\n",
+      "[0.810466468334198,],\n",
+      "[-0.9385530948638916,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.33206847310066223,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.0534361600875854,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.657759964466095,],\n",
+      "[-0.13257889449596405,],\n",
+      "[0.5354541540145874,],\n",
+      "[-0.997926652431488,],\n",
+      "[-0.08563733845949173,],\n",
+      "[0.44728773832321167,],\n",
+      "[-0.4973960220813751,],\n",
+      "[0.925058901309967,],\n",
+      "[0.810466468334198,],\n",
+      "[-0.9385530948638916,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[6.821822898928076e-05,],\n",
+      "[-2.671111906238366e-05,],\n",
+      "[-0.00048023651470430195,],\n",
+      "[0.0001398507010890171,],\n",
+      "[-0.41128838062286377,],\n",
+      "[-0.05229019746184349,],\n",
+      "[-0.5688085556030273,],\n",
+      "[-0.3653612732887268,],\n",
+      "[-0.09797253459692001,],\n",
+      "[-0.001436653663404286,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.33206847310066223,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.052738666534424,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577599048614502,],\n",
+      "[-0.13257886469364166,],\n",
+      "[0.5354546308517456,],\n",
+      "[-0.9979267716407776,],\n",
+      "[-0.08522605150938034,],\n",
+      "[0.4473400413990021,],\n",
+      "[-0.49682721495628357,],\n",
+      "[0.9254242777824402,],\n",
+      "[0.8105644583702087,],\n",
+      "[-0.938551664352417,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.33001574873924255,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4936258792877197,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577599048614502,],\n",
+      "[-0.13257886469364166,],\n",
+      "[0.5354546308517456,],\n",
+      "[-0.9979267716407776,],\n",
+      "[-0.08522605150938034,],\n",
+      "[0.4473400413990021,],\n",
+      "[-0.49682721495628357,],\n",
+      "[0.9254242777824402,],\n",
+      "[0.8105644583702087,],\n",
+      "[-0.938551664352417,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.4508644603192806e-05,],\n",
+      "[-1.3936913092038594e-05,],\n",
+      "[-0.0002645609201863408,],\n",
+      "[7.857914897613227e-05,],\n",
+      "[-0.4429711699485779,],\n",
+      "[-0.06007383018732071,],\n",
+      "[-0.6407424807548523,],\n",
+      "[-0.3633445203304291,],\n",
+      "[-0.0979747325181961,],\n",
+      "[-0.0014366526156663895,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.33001574873924255,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4051707983016968,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354548692703247,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08478307723999023,],\n",
+      "[0.44740012288093567,],\n",
+      "[-0.4961864650249481,],\n",
+      "[0.925787627696991,],\n",
+      "[0.8106624484062195,],\n",
+      "[-0.9385502338409424,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3286105692386627,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09616813063621521,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354548692703247,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08478307723999023,],\n",
+      "[0.44740012288093567,],\n",
+      "[-0.4961864650249481,],\n",
+      "[0.925787627696991,],\n",
+      "[0.8106624484062195,],\n",
+      "[-0.9385502338409424,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.1255816389166284e-05,],\n",
+      "[-4.677227934735129e-06,],\n",
+      "[-9.392101492267102e-05,],\n",
+      "[2.846880852303002e-05,],\n",
+      "[-0.47623923420906067,],\n",
+      "[-0.06881190836429596,],\n",
+      "[-0.727536141872406,],\n",
+      "[-0.3619692623615265,],\n",
+      "[-0.0979754626750946,],\n",
+      "[-0.0014366521500051022,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3286105692386627,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6202198266983032,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354549884796143,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08430683612823486,],\n",
+      "[0.44746893644332886,],\n",
+      "[-0.4954589307308197,],\n",
+      "[0.9261496067047119,],\n",
+      "[0.8107604384422302,],\n",
+      "[-0.9385488033294678,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3279903531074524,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0069717708975076675,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354549884796143,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08430683612823486,],\n",
+      "[0.44746893644332886,],\n",
+      "[-0.4954589307308197,],\n",
+      "[0.9261496067047119,],\n",
+      "[0.8107604384422302,],\n",
+      "[-0.9385488033294678,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.2395777250494575e-06,],\n",
+      "[-9.54708298195328e-07,],\n",
+      "[-2.0325005607446656e-05,],\n",
+      "[6.291701538430061e-06,],\n",
+      "[-0.493875116109848,],\n",
+      "[-0.07360641658306122,],\n",
+      "[-0.7768409252166748,],\n",
+      "[-0.36136099696159363,],\n",
+      "[-0.0979762002825737,],\n",
+      "[-0.0014366520335897803,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3279903531074524,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.16699428856372833,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354549884796143,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08381295949220657,],\n",
+      "[0.44754254817962646,],\n",
+      "[-0.49468210339546204,],\n",
+      "[0.9265109896659851,],\n",
+      "[0.810858428478241,],\n",
+      "[-0.9385473728179932,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.32782337069511414,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.02452962100505829,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354549884796143,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08381295949220657,],\n",
+      "[0.44754254817962646,],\n",
+      "[-0.49468210339546204,],\n",
+      "[0.9265109896659851,],\n",
+      "[0.810858428478241,],\n",
+      "[-0.9385473728179932,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.1043764465721324e-06,],\n",
+      "[-1.3553070630223374e-06,],\n",
+      "[-3.062905670958571e-05,],\n",
+      "[9.6868443506537e-06,],\n",
+      "[-0.48914727568626404,],\n",
+      "[-0.072194904088974,],\n",
+      "[-0.7624011635780334,],\n",
+      "[-0.3611910343170166,],\n",
+      "[-0.0979762002825737,],\n",
+      "[-0.0014366520335897803,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.32782337069511414,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3132384717464447,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.535455048084259,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08332381397485733,],\n",
+      "[0.44761472940444946,],\n",
+      "[-0.49391970038414,],\n",
+      "[0.926872193813324,],\n",
+      "[0.8109564185142517,],\n",
+      "[-0.9385459423065186,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.32751011848449707,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.23189367353916168,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.535455048084259,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08332381397485733,],\n",
+      "[0.44761472940444946,],\n",
+      "[-0.49391970038414,],\n",
+      "[0.926872193813324,],\n",
+      "[0.8109564185142517,],\n",
+      "[-0.9385459423065186,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.053507488308242e-06,],\n",
+      "[-3.1554707220493583e-06,],\n",
+      "[-7.566878048237413e-05,],\n",
+      "[2.444648816890549e-05,],\n",
+      "[-0.468545526266098,],\n",
+      "[-0.06607392430305481,],\n",
+      "[-0.6991153359413147,],\n",
+      "[-0.3608718514442444,],\n",
+      "[-0.0979762002825737,],\n",
+      "[-0.0014366520335897803,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.32751011848449707,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.963108241558075,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354551076889038,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08285526931285858,],\n",
+      "[0.447680801153183,],\n",
+      "[-0.49322059750556946,],\n",
+      "[0.9272330403327942,],\n",
+      "[0.8110544085502625,],\n",
+      "[-0.938544511795044,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3265469968318939,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7357228994369507,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354551076889038,],\n",
+      "[-0.9979268312454224,],\n",
+      "[-0.08285526931285858,],\n",
+      "[0.447680801153183,],\n",
+      "[-0.49322059750556946,],\n",
+      "[0.9272330403327942,],\n",
+      "[0.8110544085502625,],\n",
+      "[-0.938544511795044,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[9.284304724133108e-06,],\n",
+      "[-4.2662591113185044e-06,],\n",
+      "[-0.00010835991997737437,],\n",
+      "[3.574026413843967e-05,],\n",
+      "[-0.4472469389438629,],\n",
+      "[-0.0594315305352211,],\n",
+      "[-0.6278216242790222,],\n",
+      "[-0.35990282893180847,],\n",
+      "[-0.09797545522451401,],\n",
+      "[-0.0014366519171744585,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3265469968318939,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.715484619140625,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354552268981934,],\n",
+      "[-0.9979268908500671,],\n",
+      "[-0.08240802586078644,],\n",
+      "[0.44774022698402405,],\n",
+      "[-0.49259278178215027,],\n",
+      "[0.927592933177948,],\n",
+      "[0.8111523985862732,],\n",
+      "[-0.9385430812835693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3248315155506134,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1099467277526855,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354552268981934,],\n",
+      "[-0.9979268908500671,],\n",
+      "[-0.08240802586078644,],\n",
+      "[0.44774022698402405,],\n",
+      "[-0.49259278178215027,],\n",
+      "[0.927592933177948,],\n",
+      "[0.8111523985862732,],\n",
+      "[-0.9385430812835693,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[8.427005923294928e-06,],\n",
+      "[-3.988481694250368e-06,],\n",
+      "[-0.00010707731416914612,],\n",
+      "[3.603091317927465e-05,],\n",
+      "[-0.43929338455200195,],\n",
+      "[-0.056461695581674576,],\n",
+      "[-0.5900439023971558,],\n",
+      "[-0.3581903576850891,],\n",
+      "[-0.09797544777393341,],\n",
+      "[-0.0014366518007591367,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3248315155506134,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1070809364318848,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354553461074829,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08196873217821121,],\n",
+      "[0.44779670238494873,],\n",
+      "[-0.4920027256011963,],\n",
+      "[0.9279510974884033,],\n",
+      "[0.8112503886222839,],\n",
+      "[-0.9385416507720947,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.32272443175315857,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8960952758789062,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354553461074829,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08196873217821121,],\n",
+      "[0.44779670238494873,],\n",
+      "[-0.4920027256011963,],\n",
+      "[0.9279510974884033,],\n",
+      "[0.8112503886222839,],\n",
+      "[-0.9385416507720947,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[5.595380571321584e-06,],\n",
+      "[-2.7315159059071448e-06,],\n",
+      "[-7.743074820609763e-05,],\n",
+      "[2.6571937269181944e-05,],\n",
+      "[-0.4489844739437103,],\n",
+      "[-0.05881046503782272,],\n",
+      "[-0.608218789100647,],\n",
+      "[-0.3560863137245178,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.32272443175315857,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8932461738586426,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554057121277,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08151974529027939,],\n",
+      "[0.44785550236701965,],\n",
+      "[-0.4913945198059082,],\n",
+      "[0.9283071756362915,],\n",
+      "[0.8113483786582947,],\n",
+      "[-0.9385402202606201,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.32083117961883545,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3580094575881958,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554057121277,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08151974529027939,],\n",
+      "[0.44785550236701965,],\n",
+      "[-0.4913945198059082,],\n",
+      "[0.9283071756362915,],\n",
+      "[0.8113483786582947,],\n",
+      "[-0.9385402202606201,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.6135464850085555e-06,],\n",
+      "[-1.31486808641057e-06,],\n",
+      "[-3.938111331081018e-05,],\n",
+      "[1.3785212104266975e-05,],\n",
+      "[-0.46963146328926086,],\n",
+      "[-0.06482883542776108,],\n",
+      "[-0.6733090877532959,],\n",
+      "[-0.35419905185699463,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.32083117961883545,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.1966784000396729,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08105011284351349,],\n",
+      "[0.4479203224182129,],\n",
+      "[-0.490721195936203,],\n",
+      "[0.9286613464355469,],\n",
+      "[0.8114463686943054,],\n",
+      "[-0.9385387897491455,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.31963449716567993,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0469222217798233,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08105011284351349,],\n",
+      "[0.4479203224182129,],\n",
+      "[-0.490721195936203,],\n",
+      "[0.9286613464355469,],\n",
+      "[0.8114463686943054,],\n",
+      "[-0.9385387897491455,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[6.992023031671124e-07,],\n",
+      "[-3.6159499927634897e-07,],\n",
+      "[-1.146448357758345e-05,],\n",
+      "[4.096142220078036e-06,],\n",
+      "[-0.4892796576023102,],\n",
+      "[-0.07089928537607193,],\n",
+      "[-0.7461386322975159,],\n",
+      "[-0.35300537943840027,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.31963449716567993,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.43323057889938354,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.0805608332157135,],\n",
+      "[0.4479912221431732,],\n",
+      "[-0.48997506499290466,],\n",
+      "[0.9290143251419067,],\n",
+      "[0.8115443587303162,],\n",
+      "[-0.9385373592376709,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.31920126080513,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0008811806910671294,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.0805608332157135,],\n",
+      "[0.4479912221431732,],\n",
+      "[-0.48997506499290466,],\n",
+      "[0.9290143251419067,],\n",
+      "[0.8115443587303162,],\n",
+      "[-0.9385373592376709,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.080701180939286e-08,],\n",
+      "[-3.753726574018401e-08,],\n",
+      "[-1.2624622058865498e-06,],\n",
+      "[4.607217647389916e-07,],\n",
+      "[-0.49807682633399963,],\n",
+      "[-0.07370571792125702,],\n",
+      "[-0.781469464302063,],\n",
+      "[-0.3525751233100891,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.31920126080513,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.05936930328607559,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08006275445222855,],\n",
+      "[0.448064923286438,],\n",
+      "[-0.4891935884952545,],\n",
+      "[0.9293668866157532,],\n",
+      "[0.8116423487663269,],\n",
+      "[-0.9385359287261963,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3191418945789337,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.021754954010248184,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.08006275445222855,],\n",
+      "[0.448064923286438,],\n",
+      "[-0.4891935884952545,],\n",
+      "[0.9293668866157532,],\n",
+      "[0.8116423487663269,],\n",
+      "[-0.9385359287261963,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.5998755859291123e-07,],\n",
+      "[-1.410977006344183e-07,],\n",
+      "[-5.038877588958712e-06,],\n",
+      "[1.8789029354593367e-06,],\n",
+      "[-0.4932461082935333,],\n",
+      "[-0.07207569479942322,],\n",
+      "[-0.760562539100647,],\n",
+      "[-0.3525157570838928,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3191418945789337,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.2949911057949066,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.079569511115551,],\n",
+      "[0.44813698530197144,],\n",
+      "[-0.48843303322792053,],\n",
+      "[0.9297193884849548,],\n",
+      "[0.8117403388023376,],\n",
+      "[-0.9385344982147217,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3188469111919403,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.24520684778690338,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.079569511115551,],\n",
+      "[0.44813698530197144,],\n",
+      "[-0.48843303322792053,],\n",
+      "[0.9297193884849548,],\n",
+      "[0.8117403388023376,],\n",
+      "[-0.9385344982147217,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[6.450158025472774e-07,],\n",
+      "[-3.58664038913048e-07,],\n",
+      "[-1.3592059985967353e-05,],\n",
+      "[5.177388175070519e-06,],\n",
+      "[-0.47994738817214966,],\n",
+      "[-0.0674976333975792,],\n",
+      "[-0.7008385062217712,],\n",
+      "[-0.35221779346466064,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3188469111919403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.99036705493927,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.07908956706523895,],\n",
+      "[0.44820448756217957,],\n",
+      "[-0.4877322018146515,],\n",
+      "[0.9300715923309326,],\n",
+      "[0.8118383288383484,],\n",
+      "[-0.9385330677032471,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3178565502166748,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7371494174003601,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.07908956706523895,],\n",
+      "[0.44820448756217957,],\n",
+      "[-0.4877322018146515,],\n",
+      "[0.9300715923309326,],\n",
+      "[0.8118383288383484,],\n",
+      "[-0.9385330677032471,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[8.264420330306166e-07,],\n",
+      "[-4.720220090348448e-07,],\n",
+      "[-1.894695196824614e-05,],\n",
+      "[7.3678170338098425e-06,],\n",
+      "[-0.4677564203739166,],\n",
+      "[-0.06307730078697205,],\n",
+      "[-0.6389914751052856,],\n",
+      "[-0.35122743248939514,],\n",
+      "[-0.0979761853814125,],\n",
+      "[-0.0014366516843438148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3178565502166748,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7171493768692017,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.07862181216478348,],\n",
+      "[0.4482675790786743,],\n",
+      "[-0.4870932102203369,],\n",
+      "[0.930422842502594,],\n",
+      "[0.8119363188743591,],\n",
+      "[-0.9385316371917725,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3161394000053406,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [2/10], Loss: 0.7371\n",
+      "loss_antes tensor([1.1547526121139526,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6577598452568054,],\n",
+      "[-0.13257884979248047,],\n",
+      "[0.5354554653167725,],\n",
+      "[-0.9979269504547119,],\n",
+      "[-0.07862181216478348,],\n",
+      "[0.4482675790786743,],\n",
+      "[-0.4870932102203369,],\n",
+      "[0.930422842502594,],\n",
+      "[0.8119363188743591,],\n",
+      "[-0.9385316371917725,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4459612965583801,],\n",
+      "[-0.15018585324287415,],\n",
+      "[-0.541097104549408,],\n",
+      "[0.0958360880613327,],\n",
+      "[-0.23142392933368683,],\n",
+      "[-0.0360928475856781,],\n",
+      "[-0.38693270087242126,],\n",
+      "[-0.19714590907096863,],\n",
+      "[-0.05313859134912491,],\n",
+      "[-0.0008125607855618,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3161394000053406,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1491875648498535,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.657313883304596,],\n",
+      "[-0.13242866098880768,],\n",
+      "[0.5359965562820435,],\n",
+      "[-0.9980227947235107,],\n",
+      "[-0.07839038968086243,],\n",
+      "[0.4483036696910858,],\n",
+      "[-0.48670628666877747,],\n",
+      "[0.930620014667511,],\n",
+      "[0.8119894862174988,],\n",
+      "[-0.9385308027267456,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.31399020552635193,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1643296480178833,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.657313883304596,],\n",
+      "[-0.13242866098880768,],\n",
+      "[0.5359965562820435,],\n",
+      "[-0.9980227947235107,],\n",
+      "[-0.07839038968086243,],\n",
+      "[0.4483036696910858,],\n",
+      "[-0.48670628666877747,],\n",
+      "[0.930620014667511,],\n",
+      "[0.8119894862174988,],\n",
+      "[-0.9385308027267456,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.48192161321640015,],\n",
+      "[-0.15119613707065582,],\n",
+      "[-0.5080284476280212,],\n",
+      "[0.08697306364774704,],\n",
+      "[-0.23076695203781128,],\n",
+      "[-0.034878209233284,],\n",
+      "[-0.38050681352615356,],\n",
+      "[-0.21143601834774017,],\n",
+      "[-0.04738764092326164,],\n",
+      "[-0.0005327523103915155,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.31399020552635193,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1580827236175537,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6568319797515869,],\n",
+      "[-0.1322774589061737,],\n",
+      "[0.5365045666694641,],\n",
+      "[-0.998109757900238,],\n",
+      "[-0.0781596228480339,],\n",
+      "[0.4483385384082794,],\n",
+      "[-0.48632577061653137,],\n",
+      "[0.9308314323425293,],\n",
+      "[0.8120368719100952,],\n",
+      "[-0.9385302662849426,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.31183212995529175,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.6755621433258057,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6568319797515869,],\n",
+      "[-0.1322774589061737,],\n",
+      "[0.5365045666694641,],\n",
+      "[-0.998109757900238,],\n",
+      "[-0.0781596228480339,],\n",
+      "[0.4483385384082794,],\n",
+      "[-0.48632577061653137,],\n",
+      "[0.9308314323425293,],\n",
+      "[0.8120368719100952,],\n",
+      "[-0.9385302662849426,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5952900052070618,],\n",
+      "[-0.1750331073999405,],\n",
+      "[-0.5580403804779053,],\n",
+      "[0.09302129596471786,],\n",
+      "[-0.17875482141971588,],\n",
+      "[-0.02643887884914875,],\n",
+      "[-0.29498788714408875,],\n",
+      "[-0.20295339822769165,],\n",
+      "[-0.03396056964993477,],\n",
+      "[-0.00024817188386805356,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.31183212995529175,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.5888686180114746,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6562367081642151,],\n",
+      "[-0.13210242986679077,],\n",
+      "[0.5370625853538513,],\n",
+      "[-0.9982028007507324,],\n",
+      "[-0.07798086851835251,],\n",
+      "[0.4483649730682373,],\n",
+      "[-0.48603078722953796,],\n",
+      "[0.9310343861579895,],\n",
+      "[0.8120708465576172,],\n",
+      "[-0.9385300278663635,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.30924326181411743,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.2598841190338135,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6562367081642151,],\n",
+      "[-0.13210242986679077,],\n",
+      "[0.5370625853538513,],\n",
+      "[-0.9982028007507324,],\n",
+      "[-0.07798086851835251,],\n",
+      "[0.4483649730682373,],\n",
+      "[-0.48603078722953796,],\n",
+      "[0.9310343861579895,],\n",
+      "[0.8120708465576172,],\n",
+      "[-0.9385300278663635,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.679948091506958,],\n",
+      "[-0.18925417959690094,],\n",
+      "[-0.5822004079818726,],\n",
+      "[0.09511342644691467,],\n",
+      "[-0.12980739772319794,],\n",
+      "[-0.01837293803691864,],\n",
+      "[-0.21112892031669617,],\n",
+      "[-0.20108160376548767,],\n",
+      "[-0.023767350241541862,],\n",
+      "[1.8464837921783328e-05,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.30924326181411743,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-3.0065791606903076,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6555567383766174,],\n",
+      "[-0.1319131702184677,],\n",
+      "[0.5376448035240173,],\n",
+      "[-0.998297929763794,],\n",
+      "[-0.0778510645031929,],\n",
+      "[0.4483833312988281,],\n",
+      "[-0.48581966757774353,],\n",
+      "[0.9312354922294617,],\n",
+      "[0.8120946288108826,],\n",
+      "[-0.9385300278663635,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.3062366843223572,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.234950304031372,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6555567383766174,],\n",
+      "[-0.1319131702184677,],\n",
+      "[0.5376448035240173,],\n",
+      "[-0.998297929763794,],\n",
+      "[-0.0778510645031929,],\n",
+      "[0.4483833312988281,],\n",
+      "[-0.48581966757774353,],\n",
+      "[0.9312354922294617,],\n",
+      "[0.8120946288108826,],\n",
+      "[-0.9385300278663635,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.6354647874832153,],\n",
+      "[-0.16951806843280792,],\n",
+      "[-0.5111573338508606,],\n",
+      "[0.08236122876405716,],\n",
+      "[-0.13492316007614136,],\n",
+      "[-0.017848758026957512,],\n",
+      "[-0.20974138379096985,],\n",
+      "[-0.22121405601501465,],\n",
+      "[-0.027343163266777992,],\n",
+      "[0.00028353638481348753,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.3062366843223572,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.9899518489837646,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.654921293258667,],\n",
+      "[-0.1317436546087265,],\n",
+      "[0.5381559729576111,],\n",
+      "[-0.9983803033828735,],\n",
+      "[-0.07771614193916321,],\n",
+      "[0.44840118288993835,],\n",
+      "[-0.48560991883277893,],\n",
+      "[0.9314566850662231,],\n",
+      "[0.8121219873428345,],\n",
+      "[-0.9385303258895874,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.30324673652648926,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.4292161464691162,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.654921293258667,],\n",
+      "[-0.1317436546087265,],\n",
+      "[0.5381559729576111,],\n",
+      "[-0.9983803033828735,],\n",
+      "[-0.07771614193916321,],\n",
+      "[0.44840118288993835,],\n",
+      "[-0.48560991883277893,],\n",
+      "[0.9314566850662231,],\n",
+      "[0.8121219873428345,],\n",
+      "[-0.9385303258895874,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4574415981769562,],\n",
+      "[-0.11842028051614761,],\n",
+      "[-0.355233758687973,],\n",
+      "[0.056815896183252335,],\n",
+      "[-0.20979496836662292,],\n",
+      "[-0.026995470747351646,],\n",
+      "[-0.32067424058914185,],\n",
+      "[-0.25857704877853394,],\n",
+      "[-0.044902339577674866,],\n",
+      "[0.0005251052207313478,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.30324673652648926,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.3909969329833984,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6544638276100159,],\n",
+      "[-0.131625235080719,],\n",
+      "[0.5385112166404724,],\n",
+      "[-0.9984371066093445,],\n",
+      "[-0.07750634849071503,],\n",
+      "[0.4484281837940216,],\n",
+      "[-0.48528924584388733,],\n",
+      "[0.9317152500152588,],\n",
+      "[0.8121668696403503,],\n",
+      "[-0.9385308623313904,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.30085572600364685,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5100699067115784,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6544638276100159,],\n",
+      "[-0.131625235080719,],\n",
+      "[0.5385112166404724,],\n",
+      "[-0.9984371066093445,],\n",
+      "[-0.07750634849071503,],\n",
+      "[0.4484281837940216,],\n",
+      "[-0.48528924584388733,],\n",
+      "[0.9317152500152588,],\n",
+      "[0.8121668696403503,],\n",
+      "[-0.9385308623313904,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.23664869368076324,],\n",
+      "[-0.0600866936147213,],\n",
+      "[-0.18186870217323303,],\n",
+      "[0.029054244980216026,],\n",
+      "[-0.3256881535053253,],\n",
+      "[-0.041901469230651855,],\n",
+      "[-0.5031391978263855,],\n",
+      "[-0.297151654958725,],\n",
+      "[-0.06682851165533066,],\n",
+      "[0.0006954799173399806,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.30085572600364685,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4283818006515503,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6542271971702576,],\n",
+      "[-0.1315651535987854,],\n",
+      "[0.5386930704116821,],\n",
+      "[-0.9984661340713501,],\n",
+      "[-0.0771806612610817,],\n",
+      "[0.4484700858592987,],\n",
+      "[-0.48478609323501587,],\n",
+      "[0.9320123791694641,],\n",
+      "[0.8122336864471436,],\n",
+      "[-0.9385315775871277,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.299427330493927,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07758558541536331,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6542271971702576,],\n",
+      "[-0.1315651535987854,],\n",
+      "[0.5386930704116821,],\n",
+      "[-0.9984661340713501,],\n",
+      "[-0.0771806612610817,],\n",
+      "[0.4484700858592987,],\n",
+      "[-0.48478609323501587,],\n",
+      "[0.9320123791694641,],\n",
+      "[0.8122336864471436,],\n",
+      "[-0.9385315775871277,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.07731616497039795,],\n",
+      "[-0.019411372020840645,],\n",
+      "[-0.060066960752010345,],\n",
+      "[0.009637388400733471,],\n",
+      "[-0.42800143361091614,],\n",
+      "[-0.05533931776881218,],\n",
+      "[-0.6698088049888611,],\n",
+      "[-0.3229348659515381,],\n",
+      "[-0.08253288269042969,],\n",
+      "[0.0007790765375830233,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.299427330493927,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5570830702781677,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6541498899459839,],\n",
+      "[-0.13154573738574982,],\n",
+      "[0.5387531518936157,],\n",
+      "[-0.9984757900238037,],\n",
+      "[-0.0767526626586914,],\n",
+      "[0.44852542877197266,],\n",
+      "[-0.4841162860393524,],\n",
+      "[0.9323353171348572,],\n",
+      "[0.8123162388801575,],\n",
+      "[-0.9385323524475098,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2988702356815338,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.009351206943392754,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6541498899459839,],\n",
+      "[-0.13154573738574982,],\n",
+      "[0.5387531518936157,],\n",
+      "[-0.9984757900238037,],\n",
+      "[-0.0767526626586914,],\n",
+      "[0.44852542877197266,],\n",
+      "[-0.4841162860393524,],\n",
+      "[0.9323353171348572,],\n",
+      "[0.8123162388801575,],\n",
+      "[-0.9385323524475098,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.021875645965337753,],\n",
+      "[-0.005466102156788111,],\n",
+      "[-0.017486248165369034,],\n",
+      "[0.0028298699762672186,],\n",
+      "[-0.46996772289276123,],\n",
+      "[-0.06089907884597778,],\n",
+      "[-0.739665687084198,],\n",
+      "[-0.33158618211746216,],\n",
+      "[-0.08793479204177856,],\n",
+      "[0.0008029410964809358,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2988702356815338,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.19340354204177856,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6541280150413513,],\n",
+      "[-0.13154026865959167,],\n",
+      "[0.5387706160545349,],\n",
+      "[-0.9984785914421082,],\n",
+      "[-0.07628269493579865,],\n",
+      "[0.4485863149166107,],\n",
+      "[-0.4833766222000122,],\n",
+      "[0.9326668977737427,],\n",
+      "[0.8124041557312012,],\n",
+      "[-0.9385331273078918,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.29867681860923767,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.05343202129006386,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6541280150413513,],\n",
+      "[-0.13154026865959167,],\n",
+      "[0.5387706160545349,],\n",
+      "[-0.9984785914421082,],\n",
+      "[-0.07628269493579865,],\n",
+      "[0.4485863149166107,],\n",
+      "[-0.4833766222000122,],\n",
+      "[0.9326668977737427,],\n",
+      "[0.8124041557312012,],\n",
+      "[-0.9385331273078918,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.041663508862257004,],\n",
+      "[-0.010428017005324364,],\n",
+      "[-0.03476610779762268,],\n",
+      "[0.0056931315921247005,],\n",
+      "[-0.44055837392807007,],\n",
+      "[-0.05651368200778961,],\n",
+      "[-0.6873251795768738,],\n",
+      "[-0.3279472589492798,],\n",
+      "[-0.08605879545211792,],\n",
+      "[0.0008046597940847278,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.29867681860923767,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.462307870388031,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6540863513946533,],\n",
+      "[-0.13152983784675598,],\n",
+      "[0.538805365562439,],\n",
+      "[-0.9984843134880066,],\n",
+      "[-0.07584213465452194,],\n",
+      "[0.4486428201198578,],\n",
+      "[-0.4826892912387848,],\n",
+      "[0.9329948425292969,],\n",
+      "[0.8124902248382568,],\n",
+      "[-0.9385339021682739,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.29821452498435974,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.313955694437027,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6540863513946533,],\n",
+      "[-0.13152983784675598,],\n",
+      "[0.538805365562439,],\n",
+      "[-0.9984843134880066,],\n",
+      "[-0.07584213465452194,],\n",
+      "[0.4486428201198578,],\n",
+      "[-0.4826892912387848,],\n",
+      "[0.9329948425292969,],\n",
+      "[0.8124902248382568,],\n",
+      "[-0.9385339021682739,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.07897919416427612,],\n",
+      "[-0.019946325570344925,],\n",
+      "[-0.06962048262357712,],\n",
+      "[0.01156323216855526,],\n",
+      "[-0.37019190192222595,],\n",
+      "[-0.04587123543024063,],\n",
+      "[-0.5592024326324463,],\n",
+      "[-0.3206497132778168,],\n",
+      "[-0.08260372281074524,],\n",
+      "[0.0008073526551015675,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.29821452498435974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.1206347942352295,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6540073752403259,],\n",
+      "[-0.13150988519191742,],\n",
+      "[0.5388749837875366,],\n",
+      "[-0.998495876789093,],\n",
+      "[-0.07547194510698318,],\n",
+      "[0.44868868589401245,],\n",
+      "[-0.4821300804615021,],\n",
+      "[0.9333155155181885,],\n",
+      "[0.8125728368759155,],\n",
+      "[-0.9385347366333008,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2970938980579376,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7251137495040894,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6540073752403259,],\n",
+      "[-0.13150988519191742,],\n",
+      "[0.5388749837875366,],\n",
+      "[-0.998495876789093,],\n",
+      "[-0.07547194510698318,],\n",
+      "[0.44868868589401245,],\n",
+      "[-0.4821300804615021,],\n",
+      "[0.9333155155181885,],\n",
+      "[0.8125728368759155,],\n",
+      "[-0.9385347366333008,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.09245047718286514,],\n",
+      "[-0.023737872019410133,],\n",
+      "[-0.08691002428531647,],\n",
+      "[0.014667685143649578,],\n",
+      "[-0.3111974895000458,],\n",
+      "[-0.03640257567167282,],\n",
+      "[-0.4457750618457794,],\n",
+      "[-0.3164253532886505,],\n",
+      "[-0.08143877238035202,],\n",
+      "[0.000816846382804215,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2970938980579376,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7030718326568604,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6539149284362793,],\n",
+      "[-0.13148614764213562,],\n",
+      "[0.5389618873596191,],\n",
+      "[-0.9985105395317078,],\n",
+      "[-0.07516074925661087,],\n",
+      "[0.4487250745296478,],\n",
+      "[-0.48168429732322693,],\n",
+      "[0.933631956577301,],\n",
+      "[0.8126542568206787,],\n",
+      "[-0.9385355710983276,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2953908145427704,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8124265074729919,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6539149284362793,],\n",
+      "[-0.13148614764213562,],\n",
+      "[0.5389618873596191,],\n",
+      "[-0.9985105395317078,],\n",
+      "[-0.07516074925661087,],\n",
+      "[0.4487250745296478,],\n",
+      "[-0.48168429732322693,],\n",
+      "[0.933631956577301,],\n",
+      "[0.8126542568206787,],\n",
+      "[-0.9385355710983276,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.07449088990688324,],\n",
+      "[-0.019563719630241394,],\n",
+      "[-0.07520615309476852,],\n",
+      "[0.01291725691407919,],\n",
+      "[-0.30529043078422546,],\n",
+      "[-0.03453805297613144,],\n",
+      "[-0.42623013257980347,],\n",
+      "[-0.31735309958457947,],\n",
+      "[-0.08312699943780899,],\n",
+      "[0.0008295192383229733,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2953908145427704,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8026946783065796,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6538404226303101,],\n",
+      "[-0.1314665824174881,],\n",
+      "[0.5390371084213257,],\n",
+      "[-0.998523473739624,],\n",
+      "[-0.07485546171665192,],\n",
+      "[0.44875961542129517,],\n",
+      "[-0.4812580645084381,],\n",
+      "[0.9339492917060852,],\n",
+      "[0.8127374053001404,],\n",
+      "[-0.9385364055633545,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.29358813166618347,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4330214560031891,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6538404226303101,],\n",
+      "[-0.1314665824174881,],\n",
+      "[0.5390371084213257,],\n",
+      "[-0.998523473739624,],\n",
+      "[-0.07485546171665192,],\n",
+      "[0.44875961542129517,],\n",
+      "[-0.4812580645084381,],\n",
+      "[0.9339492917060852,],\n",
+      "[0.8127374053001404,],\n",
+      "[-0.9385364055633545,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04104474186897278,],\n",
+      "[-0.01106014009565115,],\n",
+      "[-0.04471351206302643,],\n",
+      "[0.007826319895684719,],\n",
+      "[-0.3587712049484253,],\n",
+      "[-0.04162505269050598,],\n",
+      "[-0.520211398601532,],\n",
+      "[-0.32163622975349426,],\n",
+      "[-0.08612129092216492,],\n",
+      "[0.0008402251987718046,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.29358813166618347,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.3160866498947144,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537993550300598,],\n",
+      "[-0.13145552575588226,],\n",
+      "[0.5390818119049072,],\n",
+      "[-0.9985312819480896,],\n",
+      "[-0.07449669390916824,],\n",
+      "[0.4488012492656708,],\n",
+      "[-0.4807378649711609,],\n",
+      "[0.9342709183692932,],\n",
+      "[0.8128235340118408,],\n",
+      "[-0.9385372400283813,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.29227203130722046,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.06360384076833725,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537993550300598,],\n",
+      "[-0.13145552575588226,],\n",
+      "[0.5390818119049072,],\n",
+      "[-0.9985312819480896,],\n",
+      "[-0.07449669390916824,],\n",
+      "[0.4488012492656708,],\n",
+      "[-0.4807378649711609,],\n",
+      "[0.9342709183692932,],\n",
+      "[0.8128235340118408,],\n",
+      "[-0.9385372400283813,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.011802807450294495,],\n",
+      "[-0.003263123333454132,],\n",
+      "[-0.013912933878600597,],\n",
+      "[0.002484481781721115,],\n",
+      "[-0.4409579336643219,],\n",
+      "[-0.053462088108062744,],\n",
+      "[-0.6766223907470703,],\n",
+      "[-0.32623234391212463,],\n",
+      "[-0.08866378664970398,],\n",
+      "[0.0008463793783448637,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.29227203130722046,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5043955445289612,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537875533103943,],\n",
+      "[-0.1314522624015808,],\n",
+      "[0.5390956997871399,],\n",
+      "[-0.9985337853431702,],\n",
+      "[-0.0740557387471199,],\n",
+      "[0.4488547146320343,],\n",
+      "[-0.48006123304367065,],\n",
+      "[0.9345971345901489,],\n",
+      "[0.8129122257232666,],\n",
+      "[-0.9385380744934082,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2917676270008087,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.00695335678756237,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537875533103943,],\n",
+      "[-0.1314522624015808,],\n",
+      "[0.5390956997871399,],\n",
+      "[-0.9985337853431702,],\n",
+      "[-0.0740557387471199,],\n",
+      "[0.4488547146320343,],\n",
+      "[-0.48006123304367065,],\n",
+      "[0.9345971345901489,],\n",
+      "[0.8129122257232666,],\n",
+      "[-0.9385380744934082,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.002916180295869708,],\n",
+      "[0.0008257061126641929,],\n",
+      "[0.0037256379146128893,],\n",
+      "[-0.0006794113432988524,],\n",
+      "[-0.5052439570426941,],\n",
+      "[-0.06319974362850189,],\n",
+      "[-0.8055168986320496,],\n",
+      "[-0.32902616262435913,],\n",
+      "[-0.08990766853094101,],\n",
+      "[0.0008484156569465995,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2917676270008087,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.16677366197109222,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537904739379883,],\n",
+      "[-0.13145308196544647,],\n",
+      "[0.539091944694519,],\n",
+      "[-0.9985331296920776,],\n",
+      "[-0.07355049252510071,],\n",
+      "[0.4489179253578186,],\n",
+      "[-0.47925570607185364,],\n",
+      "[0.9349261522293091,],\n",
+      "[0.8130021095275879,],\n",
+      "[-0.9385389089584351,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.29193440079689026,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.024249816313385963,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537904739379883,],\n",
+      "[-0.13145308196544647,],\n",
+      "[0.539091944694519,],\n",
+      "[-0.9985331296920776,],\n",
+      "[-0.07355049252510071,],\n",
+      "[0.4489179253578186,],\n",
+      "[-0.47925570607185364,],\n",
+      "[0.9349261522293091,],\n",
+      "[0.8130021095275879,],\n",
+      "[-0.9385389089584351,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.004057010635733604,],\n",
+      "[0.0011745953233912587,],\n",
+      "[0.00562351243570447,],\n",
+      "[-0.001048050937242806,],\n",
+      "[-0.5179282426834106,],\n",
+      "[-0.0653490349650383,],\n",
+      "[-0.8331149816513062,],\n",
+      "[-0.3295305669307709,],\n",
+      "[-0.08999835699796677,],\n",
+      "[0.0008482411503791809,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.29193440079689026,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3114471435546875,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.653794527053833,],\n",
+      "[-0.1314542591571808,],\n",
+      "[0.5390863418579102,],\n",
+      "[-0.9985320568084717,],\n",
+      "[-0.07303256541490555,],\n",
+      "[0.4489832818508148,],\n",
+      "[-0.4784225821495056,],\n",
+      "[0.9352557063102722,],\n",
+      "[0.8130921125411987,],\n",
+      "[-0.9385397434234619,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2922458350658417,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.004975505638867617,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.653794527053833,],\n",
+      "[-0.1314542591571808,],\n",
+      "[0.5390863418579102,],\n",
+      "[-0.9985320568084717,],\n",
+      "[-0.07303256541490555,],\n",
+      "[0.4489832818508148,],\n",
+      "[-0.4784225821495056,],\n",
+      "[0.9352557063102722,],\n",
+      "[0.8130921125411987,],\n",
+      "[-0.9385397434234619,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0013654670910909772,],\n",
+      "[-0.0004042908549308777,],\n",
+      "[-0.002055488293990493,],\n",
+      "[0.00039168517105281353,],\n",
+      "[-0.4773414433002472,],\n",
+      "[-0.05891046300530434,],\n",
+      "[-0.7469502687454224,],\n",
+      "[-0.3284381330013275,],\n",
+      "[-0.08956029266119003,],\n",
+      "[0.000847630319185555,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2922458350658417,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.14107462763786316,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537931561470032,],\n",
+      "[-0.13145385682582855,],\n",
+      "[0.5390883684158325,],\n",
+      "[-0.9985324740409851,],\n",
+      "[-0.07255522161722183,],\n",
+      "[0.4490422010421753,],\n",
+      "[-0.4776756465435028,],\n",
+      "[0.9355841279029846,],\n",
+      "[0.8131816983222961,],\n",
+      "[-0.9385405778884888,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2921047508716583,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20348051190376282,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537931561470032,],\n",
+      "[-0.13145385682582855,],\n",
+      "[0.5390883684158325,],\n",
+      "[-0.9985324740409851,],\n",
+      "[-0.07255522161722183,],\n",
+      "[0.4490422010421753,],\n",
+      "[-0.4776756465435028,],\n",
+      "[0.9355841279029846,],\n",
+      "[0.8131816983222961,],\n",
+      "[-0.9385405778884888,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.006473982241004705,],\n",
+      "[-0.0019644531421363354,],\n",
+      "[-0.010594251565635204,],\n",
+      "[0.0020644625183194876,],\n",
+      "[-0.4132333993911743,],\n",
+      "[-0.04808065667748451,],\n",
+      "[-0.6032649874687195,],\n",
+      "[-0.32676637172698975,],\n",
+      "[-0.08916120231151581,],\n",
+      "[0.0008474586647935212,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2921047508716583,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.902176022529602,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537866592407227,],\n",
+      "[-0.13145188987255096,],\n",
+      "[0.5390989780426025,],\n",
+      "[-0.9985345602035522,],\n",
+      "[-0.07214199006557465,],\n",
+      "[0.44909027218818665,],\n",
+      "[-0.4770723879337311,],\n",
+      "[0.9359108805656433,],\n",
+      "[0.8132708668708801,],\n",
+      "[-0.9385414123535156,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.291202574968338,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.561247706413269,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537866592407227,],\n",
+      "[-0.13145188987255096,],\n",
+      "[0.5390989780426025,],\n",
+      "[-0.9985345602035522,],\n",
+      "[-0.07214199006557465,],\n",
+      "[0.44909027218818665,],\n",
+      "[-0.4770723879337311,],\n",
+      "[0.9359108805656433,],\n",
+      "[0.8132708668708801,],\n",
+      "[-0.9385414123535156,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.007957303896546364,],\n",
+      "[-0.00248211482539773,],\n",
+      "[-0.01416866946965456,],\n",
+      "[0.0028232114855200052,],\n",
+      "[-0.3674349784851074,],\n",
+      "[-0.03971531614661217,],\n",
+      "[-0.4919159710407257,],\n",
+      "[-0.32525908946990967,],\n",
+      "[-0.08905268460512161,],\n",
+      "[0.0008478082017973065,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.291202574968338,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4983294010162354,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537786722183228,],\n",
+      "[-0.1314494013786316,],\n",
+      "[0.5391131639480591,],\n",
+      "[-0.9985373616218567,],\n",
+      "[-0.07177455723285675,],\n",
+      "[0.44912999868392944,],\n",
+      "[-0.4765804708003998,],\n",
+      "[0.9362361431121826,],\n",
+      "[0.8133599162101746,],\n",
+      "[-0.9385422468185425,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28970423340797424,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6068801283836365,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537786722183228,],\n",
+      "[-0.1314494013786316,],\n",
+      "[0.5391131639480591,],\n",
+      "[-0.9985373616218567,],\n",
+      "[-0.07177455723285675,],\n",
+      "[0.44912999868392944,],\n",
+      "[-0.4765804708003998,],\n",
+      "[0.9362361431121826,],\n",
+      "[0.8133599162101746,],\n",
+      "[-0.9385422468185425,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.006116185337305069,],\n",
+      "[-0.0019656929653137922,],\n",
+      "[-0.011857039295136929,],\n",
+      "[0.0024155504070222378,],\n",
+      "[-0.36766135692596436,],\n",
+      "[-0.03889438882470131,],\n",
+      "[-0.4815264642238617,],\n",
+      "[-0.32420656085014343,],\n",
+      "[-0.08919709175825119,],\n",
+      "[0.0008483699057251215,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28970423340797424,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.5580506324768066,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537725329399109,],\n",
+      "[-0.131447434425354,],\n",
+      "[0.5391250252723694,],\n",
+      "[-0.9985398054122925,],\n",
+      "[-0.07140689343214035,],\n",
+      "[0.4491688907146454,],\n",
+      "[-0.4760989546775818,],\n",
+      "[0.936560332775116,],\n",
+      "[0.8134490847587585,],\n",
+      "[-0.9385430812835693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28814616799354553,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2689037322998047,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537725329399109,],\n",
+      "[-0.131447434425354,],\n",
+      "[0.5391250252723694,],\n",
+      "[-0.9985398054122925,],\n",
+      "[-0.07140689343214035,],\n",
+      "[0.4491688907146454,],\n",
+      "[-0.4760989546775818,],\n",
+      "[0.936560332775116,],\n",
+      "[0.8134490847587585,],\n",
+      "[-0.9385430812835693,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0030070561915636063,],\n",
+      "[-0.000996038201265037,],\n",
+      "[-0.006347954738885164,],\n",
+      "[0.0013222962152212858,],\n",
+      "[-0.4116136133670807,],\n",
+      "[-0.045795541256666183,],\n",
+      "[-0.5788933038711548,],\n",
+      "[-0.32363539934158325,],\n",
+      "[-0.08942962437868118,],\n",
+      "[0.0008488429011777043,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28814616799354553,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0371193885803223,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537695527076721,],\n",
+      "[-0.13144643604755402,],\n",
+      "[0.5391314029693604,],\n",
+      "[-0.9985411167144775,],\n",
+      "[-0.0709952786564827,],\n",
+      "[0.4492146968841553,],\n",
+      "[-0.4755200743675232,],\n",
+      "[0.9368839859962463,],\n",
+      "[0.8135384917259216,],\n",
+      "[-0.9385439157485962,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28710904717445374,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.015030542388558388,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537695527076721,],\n",
+      "[-0.13144643604755402,],\n",
+      "[0.5391314029693604,],\n",
+      "[-0.9985411167144775,],\n",
+      "[-0.0709952786564827,],\n",
+      "[0.4492146968841553,],\n",
+      "[-0.4755200743675232,],\n",
+      "[0.9368839859962463,],\n",
+      "[0.8135384917259216,],\n",
+      "[-0.9385439157485962,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0005248838569968939,],\n",
+      "[-0.00017886824207380414,],\n",
+      "[-0.0012062015011906624,],\n",
+      "[0.0002569940115790814,],\n",
+      "[-0.4712323546409607,],\n",
+      "[-0.05598163604736328,],\n",
+      "[-0.7255333662033081,],\n",
+      "[-0.32349276542663574,],\n",
+      "[-0.08961043506860733,],\n",
+      "[0.0008490945328958333,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28710904717445374,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.24519853293895721,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537690162658691,],\n",
+      "[-0.1314462572336197,],\n",
+      "[0.5391325950622559,],\n",
+      "[-0.9985413551330566,],\n",
+      "[-0.07052404433488846,],\n",
+      "[0.44927066564559937,],\n",
+      "[-0.47479453682899475,],\n",
+      "[0.9372074604034424,],\n",
+      "[0.813628077507019,],\n",
+      "[-0.938544750213623,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28686386346817017,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.029305152595043182,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537690162658691,],\n",
+      "[-0.1314462572336197,],\n",
+      "[0.5391325950622559,],\n",
+      "[-0.9985413551330566,],\n",
+      "[-0.07052404433488846,],\n",
+      "[0.44927066564559937,],\n",
+      "[-0.47479453682899475,],\n",
+      "[0.9372074604034424,],\n",
+      "[0.813628077507019,],\n",
+      "[-0.938544750213623,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0005409587756730616,],\n",
+      "[0.00018915344844572246,],\n",
+      "[0.0013525483664125204,],\n",
+      "[-0.0002949122863356024,],\n",
+      "[-0.511934757232666,],\n",
+      "[-0.06336940079927444,],\n",
+      "[-0.8332101702690125,],\n",
+      "[-0.32368332147598267,],\n",
+      "[-0.08968600630760193,],\n",
+      "[0.0008491532644256949,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28686386346817017,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3423750102519989,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537695527076721,],\n",
+      "[-0.1314464509487152,],\n",
+      "[0.539131224155426,],\n",
+      "[-0.9985410571098328,],\n",
+      "[-0.07001210749149323,],\n",
+      "[0.4493340253829956,],\n",
+      "[-0.47396132349967957,],\n",
+      "[0.9375311732292175,],\n",
+      "[0.813717782497406,],\n",
+      "[-0.9385455846786499,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28720623254776,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03487062826752663,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537695527076721,],\n",
+      "[-0.1314464509487152,],\n",
+      "[0.539131224155426,],\n",
+      "[-0.9985410571098328,],\n",
+      "[-0.07001210749149323,],\n",
+      "[0.4493340253829956,],\n",
+      "[-0.47396132349967957,],\n",
+      "[0.9375311732292175,],\n",
+      "[0.813717782497406,],\n",
+      "[-0.9385455846786499,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0004354492702987045,],\n",
+      "[0.0001559262309456244,],\n",
+      "[0.0011840524384751916,],\n",
+      "[-0.0002643199695739895,],\n",
+      "[-0.5130350589752197,],\n",
+      "[-0.06377516686916351,],\n",
+      "[-0.838373064994812,],\n",
+      "[-0.3239951729774475,],\n",
+      "[-0.08967823535203934,],\n",
+      "[0.0008491246262565255,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28720623254776,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.37347373366355896,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537699699401855,],\n",
+      "[-0.13144659996032715,],\n",
+      "[0.5391300320625305,],\n",
+      "[-0.9985408186912537,],\n",
+      "[-0.06949907541275024,],\n",
+      "[0.4493978023529053,],\n",
+      "[-0.4731229543685913,],\n",
+      "[0.9378551840782166,],\n",
+      "[0.813807487487793,],\n",
+      "[-0.9385464191436768,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2875797152519226,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.006948475260287523,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537699699401855,],\n",
+      "[-0.13144659996032715,],\n",
+      "[0.5391300320625305,],\n",
+      "[-0.9985408186912537,],\n",
+      "[-0.06949907541275024,],\n",
+      "[0.4493978023529053,],\n",
+      "[-0.4731229543685913,],\n",
+      "[0.9378551840782166,],\n",
+      "[0.813807487487793,],\n",
+      "[-0.9385464191436768,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00014340471534524113,],\n",
+      "[-5.258253077045083e-05,],\n",
+      "[-0.00042404347914271057,],\n",
+      "[9.692062303656712e-05,],\n",
+      "[-0.47835367918014526,],\n",
+      "[-0.05725870281457901,],\n",
+      "[-0.7412477731704712,],\n",
+      "[-0.32409682869911194,],\n",
+      "[-0.0896390825510025,],\n",
+      "[0.0008490901091136038,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2875797152519226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.16671516001224518,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.653769850730896,],\n",
+      "[-0.13144654035568237,],\n",
+      "[0.539130449295044,],\n",
+      "[-0.9985409379005432,],\n",
+      "[-0.06902071833610535,],\n",
+      "[0.44945505261421204,],\n",
+      "[-0.47238171100616455,],\n",
+      "[0.9381792545318604,],\n",
+      "[0.8138971328735352,],\n",
+      "[-0.9385472536087036,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28741300106048584,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22395463287830353,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.653769850730896,],\n",
+      "[-0.13144654035568237,],\n",
+      "[0.539130449295044,],\n",
+      "[-0.9985409379005432,],\n",
+      "[-0.06902071833610535,],\n",
+      "[0.44945505261421204,],\n",
+      "[-0.47238171100616455,],\n",
+      "[0.9381792545318604,],\n",
+      "[0.8138971328735352,],\n",
+      "[-0.9385472536087036,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0006004933384247124,],\n",
+      "[-0.000225878888159059,],\n",
+      "[-0.0019316001562401652,],\n",
+      "[0.00045187128125689924,],\n",
+      "[-0.4324330687522888,],\n",
+      "[-0.04806327447295189,],\n",
+      "[-0.6036280393600464,],\n",
+      "[-0.32367998361587524,],\n",
+      "[-0.08960917592048645,],\n",
+      "[0.0008490837062709033,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28741300106048584,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.946475625038147,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537692546844482,],\n",
+      "[-0.13144631683826447,],\n",
+      "[0.5391323566436768,],\n",
+      "[-0.9985414147377014,],\n",
+      "[-0.06858828663825989,],\n",
+      "[0.4495031237602234,],\n",
+      "[-0.47177809476852417,],\n",
+      "[0.9385029077529907,],\n",
+      "[0.8139867186546326,],\n",
+      "[-0.9385480880737305,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2864665389060974,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.554157018661499,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537692546844482,],\n",
+      "[-0.13144631683826447,],\n",
+      "[0.5391323566436768,],\n",
+      "[-0.9985414147377014,],\n",
+      "[-0.06858828663825989,],\n",
+      "[0.4495031237602234,],\n",
+      "[-0.47177809476852417,],\n",
+      "[0.9385029077529907,],\n",
+      "[0.8139867186546326,],\n",
+      "[-0.9385480880737305,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0006965831271372736,],\n",
+      "[-0.0002695017319638282,],\n",
+      "[-0.0024387987796217203,],\n",
+      "[0.000583607004955411,],\n",
+      "[-0.40459758043289185,],\n",
+      "[-0.04192867502570152,],\n",
+      "[-0.509099543094635,],\n",
+      "[-0.32265403866767883,],\n",
+      "[-0.08960352092981339,],\n",
+      "[0.0008491040207445621,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2864665389060974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4888325929641724,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537685394287109,],\n",
+      "[-0.13144604861736298,],\n",
+      "[0.5391348004341125,],\n",
+      "[-0.9985420107841492,],\n",
+      "[-0.06818369030952454,],\n",
+      "[0.4495450556278229,],\n",
+      "[-0.47126898169517517,],\n",
+      "[0.9388255476951599,],\n",
+      "[0.81407630443573,],\n",
+      "[-0.9385489225387573,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2849777042865753,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5336568355560303,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537685394287109,],\n",
+      "[-0.13144604861736298,],\n",
+      "[0.5391348004341125,],\n",
+      "[-0.9985420107841492,],\n",
+      "[-0.06818369030952454,],\n",
+      "[0.4495450556278229,],\n",
+      "[-0.47126898169517517,],\n",
+      "[0.9388255476951599,],\n",
+      "[0.81407630443573,],\n",
+      "[-0.9385489225387573,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0005040432442910969,],\n",
+      "[-0.0002009318268392235,],\n",
+      "[-0.0019214239437133074,],\n",
+      "[0.0004701468860730529,],\n",
+      "[-0.4105534255504608,],\n",
+      "[-0.042415499687194824,],\n",
+      "[-0.514039933681488,],\n",
+      "[-0.3212558627128601,],\n",
+      "[-0.08961568772792816,],\n",
+      "[0.000849131727591157,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2849777042865753,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4610369205474854,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537680625915527,],\n",
+      "[-0.13144585490226746,],\n",
+      "[0.5391367077827454,],\n",
+      "[-0.9985424876213074,],\n",
+      "[-0.06777313351631165,],\n",
+      "[0.44958746433258057,],\n",
+      "[-0.4707549512386322,],\n",
+      "[0.9391468167304993,],\n",
+      "[0.8141659498214722,],\n",
+      "[-0.9385497570037842,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28351667523384094,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.19213294982910156,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537680625915527,],\n",
+      "[-0.13144585490226746,],\n",
+      "[0.5391367077827454,],\n",
+      "[-0.9985424876213074,],\n",
+      "[-0.06777313351631165,],\n",
+      "[0.44958746433258057,],\n",
+      "[-0.4707549512386322,],\n",
+      "[0.9391468167304993,],\n",
+      "[0.8141659498214722,],\n",
+      "[-0.9385497570037842,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00022299365082290024,],\n",
+      "[-9.158271859632805e-05,],\n",
+      "[-0.000925507687497884,],\n",
+      "[0.00023155726375989616,],\n",
+      "[-0.44417017698287964,],\n",
+      "[-0.048843514174222946,],\n",
+      "[-0.6156979203224182,],\n",
+      "[-0.31995970010757446,],\n",
+      "[-0.08963321149349213,],\n",
+      "[0.0008491522748954594,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28351667523384094,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8766592144966125,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537678241729736,],\n",
+      "[-0.1314457654953003,],\n",
+      "[0.5391376614570618,],\n",
+      "[-0.9985427260398865,],\n",
+      "[-0.06732895970344543,],\n",
+      "[0.449636310338974,],\n",
+      "[-0.4701392650604248,],\n",
+      "[0.9394667744636536,],\n",
+      "[0.8142555952072144,],\n",
+      "[-0.938550591468811,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2826400101184845,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0019166457932442427,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537678241729736,],\n",
+      "[-0.1314457654953003,],\n",
+      "[0.5391376614570618,],\n",
+      "[-0.9985427260398865,],\n",
+      "[-0.06732895970344543,],\n",
+      "[0.449636310338974,],\n",
+      "[-0.4701392650604248,],\n",
+      "[0.9394667744636536,],\n",
+      "[0.8142555952072144,],\n",
+      "[-0.938550591468811,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.642140887270216e-05,],\n",
+      "[-6.934716111572925e-06,],\n",
+      "[-7.416791777359322e-05,],\n",
+      "[1.8982591427629814e-05,],\n",
+      "[-0.4839555025100708,],\n",
+      "[-0.0570891909301281,],\n",
+      "[-0.7526813745498657,],\n",
+      "[-0.3192206025123596,],\n",
+      "[-0.08964646607637405,],\n",
+      "[0.0008491641492582858,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2826400101184845,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.08755900710821152,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537678241729736,],\n",
+      "[-0.1314457654953003,],\n",
+      "[0.5391377210617065,],\n",
+      "[-0.9985427260398865,],\n",
+      "[-0.06684500724077225,],\n",
+      "[0.4496934115886688,],\n",
+      "[-0.46938657760620117,],\n",
+      "[0.9397860169410706,],\n",
+      "[0.8143452405929565,],\n",
+      "[-0.9385514259338379,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.28255245089530945,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.046177662909030914,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537678241729736,],\n",
+      "[-0.1314457654953003,],\n",
+      "[0.5391377210617065,],\n",
+      "[-0.9985427260398865,],\n",
+      "[-0.06684500724077225,],\n",
+      "[0.4496934115886688,],\n",
+      "[-0.46938657760620117,],\n",
+      "[0.9397860169410706,],\n",
+      "[0.8143452405929565,],\n",
+      "[-0.9385514259338379,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.942895222688094e-05,],\n",
+      "[2.573906022007577e-05,],\n",
+      "[0.00029190091299824417,],\n",
+      "[-7.647315214853734e-05,],\n",
+      "[-0.5074343085289001,],\n",
+      "[-0.0622805655002594,],\n",
+      "[-0.8413334488868713,],\n",
+      "[-0.3191896975040436,],\n",
+      "[-0.08965092897415161,],\n",
+      "[0.0008491645567119122,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.28255245089530945,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4297794699668884,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537678837776184,],\n",
+      "[-0.13144579529762268,],\n",
+      "[0.5391374230384827,],\n",
+      "[-0.9985426664352417,],\n",
+      "[-0.06633757054805756,],\n",
+      "[0.4497556984424591,],\n",
+      "[-0.46854525804519653,],\n",
+      "[0.9401051998138428,],\n",
+      "[0.8144348859786987,],\n",
+      "[-0.9385522603988647,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2829822301864624,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03308456391096115,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537678837776184,],\n",
+      "[-0.13144579529762268,],\n",
+      "[0.5391374230384827,],\n",
+      "[-0.9985426664352417,],\n",
+      "[-0.06633757054805756,],\n",
+      "[0.4497556984424591,],\n",
+      "[-0.46854525804519653,],\n",
+      "[0.9401051998138428,],\n",
+      "[0.8144348859786987,],\n",
+      "[-0.9385522603988647,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-3.7087742384755984e-05,],\n",
+      "[1.644622716412414e-05,],\n",
+      "[0.0001980170636670664,],\n",
+      "[-5.312432040227577e-05,],\n",
+      "[-0.5036916732788086,],\n",
+      "[-0.06161092594265938,],\n",
+      "[-0.8293047547340393,],\n",
+      "[-0.31960442662239075,],\n",
+      "[-0.08964947611093521,],\n",
+      "[0.0008491631597280502,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2829822301864624,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3637828826904297,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537679433822632,],\n",
+      "[-0.13144581019878387,],\n",
+      "[0.5391372442245483,],\n",
+      "[-0.9985426068305969,],\n",
+      "[-0.06583388149738312,],\n",
+      "[0.4498172998428345,],\n",
+      "[-0.46771594882011414,],\n",
+      "[0.9404247999191284,],\n",
+      "[0.8145245313644409,],\n",
+      "[-0.9385530948638916,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.283346027135849,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.014756313525140285,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537679433822632,],\n",
+      "[-0.13144581019878387,],\n",
+      "[0.5391372442245483,],\n",
+      "[-0.9985426068305969,],\n",
+      "[-0.06583388149738312,],\n",
+      "[0.4498172998428345,],\n",
+      "[-0.46771594882011414,],\n",
+      "[0.9404247999191284,],\n",
+      "[0.8145245313644409,],\n",
+      "[-0.9385530948638916,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.8261322111357003e-05,],\n",
+      "[-8.29282453196356e-06,],\n",
+      "[-0.00010598616790957749,],\n",
+      "[2.911620140366722e-05,],\n",
+      "[-0.4783271253108978,],\n",
+      "[-0.05585164949297905,],\n",
+      "[-0.7281153202056885,],\n",
+      "[-0.3199204206466675,],\n",
+      "[-0.08964651823043823,],\n",
+      "[0.0008491623448207974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.283346027135849,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.2429511994123459,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537679433822632,],\n",
+      "[-0.13144579529762268,],\n",
+      "[0.5391373634338379,],\n",
+      "[-0.9985426068305969,],\n",
+      "[-0.06535555422306061,],\n",
+      "[0.449873149394989,],\n",
+      "[-0.46698784828186035,],\n",
+      "[0.9407446980476379,],\n",
+      "[0.8146141767501831,],\n",
+      "[-0.9385539293289185,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2831030786037445,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2598913908004761,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537679433822632,],\n",
+      "[-0.13144579529762268,],\n",
+      "[0.5391373634338379,],\n",
+      "[-0.9985426068305969,],\n",
+      "[-0.06535555422306061,],\n",
+      "[0.449873149394989,],\n",
+      "[-0.46698784828186035,],\n",
+      "[0.9407446980476379,],\n",
+      "[0.8146141767501831,],\n",
+      "[-0.9385539293289185,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[5.650072853313759e-05,],\n",
+      "[-2.6327883460908197e-05,],\n",
+      "[-0.00035663621383719146,],\n",
+      "[0.00010027232929132879,],\n",
+      "[-0.44908374547958374,],\n",
+      "[-0.04878808185458183,],\n",
+      "[-0.6014854311943054,],\n",
+      "[-0.31963878870010376,],\n",
+      "[-0.08964426815509796,],\n",
+      "[0.0008491626940667629,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2831030786037445,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0195908546447754,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537678837776184,],\n",
+      "[-0.1314457654953003,],\n",
+      "[0.5391377210617065,],\n",
+      "[-0.9985427260398865,],\n",
+      "[-0.06490647047758102,],\n",
+      "[0.44992193579673767,],\n",
+      "[-0.4663863480091095,],\n",
+      "[0.9410643577575684,],\n",
+      "[0.8147038221359253,],\n",
+      "[-0.9385547637939453,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2820834815502167,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5565387010574341,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537678837776184,],\n",
+      "[-0.1314457654953003,],\n",
+      "[0.5391377210617065,],\n",
+      "[-0.9985427260398865,],\n",
+      "[-0.06490647047758102,],\n",
+      "[0.44992193579673767,],\n",
+      "[-0.4663863480091095,],\n",
+      "[0.9410643577575684,],\n",
+      "[0.8147038221359253,],\n",
+      "[-0.9385547637939453,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[6.095543358242139e-05,],\n",
+      "[-2.9218788768048398e-05,],\n",
+      "[-0.0004187116282992065,],\n",
+      "[0.00012040691945003346,],\n",
+      "[-0.4343513250350952,],\n",
+      "[-0.044788796454668045,],\n",
+      "[-0.5243955850601196,],\n",
+      "[-0.3186134099960327,],\n",
+      "[-0.08964423090219498,],\n",
+      "[0.0008491637418046594,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2820834815502167,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4920321702957153,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537678241729736,],\n",
+      "[-0.1314457356929779,],\n",
+      "[0.53913813829422,],\n",
+      "[-0.998542845249176,],\n",
+      "[-0.06447211652994156,],\n",
+      "[0.4499667286872864,],\n",
+      "[-0.4658619463443756,],\n",
+      "[0.9413829445838928,],\n",
+      "[0.8147934675216675,],\n",
+      "[-0.9385555982589722,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2805914580821991,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.46957147121429443,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537678241729736,],\n",
+      "[-0.1314457356929779,],\n",
+      "[0.53913813829422,],\n",
+      "[-0.998542845249176,],\n",
+      "[-0.06447211652994156,],\n",
+      "[0.4499667286872864,],\n",
+      "[-0.4658619463443756,],\n",
+      "[0.9413829445838928,],\n",
+      "[0.8147934675216675,],\n",
+      "[-0.9385555982589722,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[4.127795182284899e-05,],\n",
+      "[-2.038503771473188e-05,],\n",
+      "[-0.00030868948670104146,],\n",
+      "[9.075227717403322e-05,],\n",
+      "[-0.4417746365070343,],\n",
+      "[-0.046066515147686005,],\n",
+      "[-0.5424510836601257,],\n",
+      "[-0.3171364665031433,],\n",
+      "[-0.08964493870735168,],\n",
+      "[0.0008491647895425558,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2805914580821991,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.3705055713653564,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.1314457207918167,],\n",
+      "[0.5391384363174438,],\n",
+      "[-0.9985429644584656,],\n",
+      "[-0.06403034180402756,],\n",
+      "[0.45001280307769775,],\n",
+      "[-0.4653194844722748,],\n",
+      "[0.9417001008987427,],\n",
+      "[0.8148831129074097,],\n",
+      "[-0.938556432723999,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27922093868255615,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13171173632144928,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.1314457207918167,],\n",
+      "[0.5391384363174438,],\n",
+      "[-0.9985429644584656,],\n",
+      "[-0.06403034180402756,],\n",
+      "[0.45001280307769775,],\n",
+      "[-0.4653194844722748,],\n",
+      "[0.9417001008987427,],\n",
+      "[0.8148831129074097,],\n",
+      "[-0.938556432723999,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.6116810002131388e-05,],\n",
+      "[-8.196640010282863e-06,],\n",
+      "[-0.00013120021321810782,],\n",
+      "[3.9437869418179616e-05,],\n",
+      "[-0.46496886014938354,],\n",
+      "[-0.0515066459774971,],\n",
+      "[-0.6464608311653137,],\n",
+      "[-0.31579291820526123,],\n",
+      "[-0.08964640647172928,],\n",
+      "[0.0008491654880344868,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27922093868255615,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7258416414260864,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391385555267334,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06356537342071533,],\n",
+      "[0.4500643014907837,],\n",
+      "[-0.4646730124950409,],\n",
+      "[0.9420158863067627,],\n",
+      "[0.8149727582931519,],\n",
+      "[-0.9385572671890259,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27849510312080383,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0007038556504994631,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391385555267334,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06356537342071533,],\n",
+      "[0.4500643014907837,],\n",
+      "[-0.4646730124950409,],\n",
+      "[0.9420158863067627,],\n",
+      "[0.8149727582931519,],\n",
+      "[-0.9385572671890259,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.685800025887147e-07,],\n",
+      "[4.539664928415732e-07,],\n",
+      "[7.692606232012622e-06,],\n",
+      "[-2.36555911214964e-06,],\n",
+      "[-0.4893682897090912,],\n",
+      "[-0.0576646625995636,],\n",
+      "[-0.7736217379570007,],\n",
+      "[-0.3150880038738251,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657208651304,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27849510312080383,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.05306050926446915,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391385555267334,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06307600438594818,],\n",
+      "[0.4501219689846039,],\n",
+      "[-0.4638994038105011,],\n",
+      "[0.9423309564590454,],\n",
+      "[0.815062403678894,],\n",
+      "[-0.9385581016540527,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27854815125465393,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.061399612575769424,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391385555267334,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06307600438594818,],\n",
+      "[0.4501219689846039,],\n",
+      "[-0.4638994038105011,],\n",
+      "[0.9423309564590454,],\n",
+      "[0.815062403678894,],\n",
+      "[-0.9385581016540527,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.980701644148212e-06,],\n",
+      "[3.204333324902109e-06,],\n",
+      "[5.7587480114307255e-05,],\n",
+      "[-1.8128597730537876e-05,],\n",
+      "[-0.5016208291053772,],\n",
+      "[-0.060970138758420944,],\n",
+      "[-0.8448084592819214,],\n",
+      "[-0.3151499927043915,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657208651304,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27854815125465393,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.49557965993881226,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384959220886,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06257438659667969,],\n",
+      "[0.4501829445362091,],\n",
+      "[-0.4630545973777771,],\n",
+      "[0.9426460862159729,],\n",
+      "[0.8151520490646362,],\n",
+      "[-0.9385589361190796,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2790437340736389,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.028161056339740753,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384959220886,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06257438659667969,],\n",
+      "[0.4501829445362091,],\n",
+      "[-0.4630545973777771,],\n",
+      "[0.9426460862159729,],\n",
+      "[0.8151520490646362,],\n",
+      "[-0.9385589361190796,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.9860307222406846e-06,],\n",
+      "[1.6377041447412921e-06,],\n",
+      "[3.124729119008407e-05,],\n",
+      "[-1.007374612527201e-05,],\n",
+      "[-0.4966736435890198,],\n",
+      "[-0.059759676456451416,],\n",
+      "[-0.8182312846183777,],\n",
+      "[-0.3156425654888153,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491656626574695,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2790437340736389,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.335625022649765,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384363174438,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06207771226763725,],\n",
+      "[0.45024269819259644,],\n",
+      "[-0.4622363746166229,],\n",
+      "[0.9429617524147034,],\n",
+      "[0.8152416944503784,],\n",
+      "[-0.9385597705841064,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27937936782836914,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.026732293888926506,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384363174438,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06207771226763725,],\n",
+      "[0.45024269819259644,],\n",
+      "[-0.4622363746166229,],\n",
+      "[0.9429617524147034,],\n",
+      "[0.8152416944503784,],\n",
+      "[-0.9385597705841064,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.1448063307616394e-06,],\n",
+      "[-1.2047731843267684e-06,],\n",
+      "[-2.4395190848736092e-05,],\n",
+      "[8.053123565332498e-06,],\n",
+      "[-0.47972342371940613,],\n",
+      "[-0.05511463060975075,],\n",
+      "[-0.7150912880897522,],\n",
+      "[-0.31596922874450684,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491656044498086,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27937936782836914,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.32700011134147644,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384363174438,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06159798800945282,],\n",
+      "[0.4502978026866913,],\n",
+      "[-0.46152129769325256,],\n",
+      "[0.9432777166366577,],\n",
+      "[0.8153313398361206,],\n",
+      "[-0.9385606050491333,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27905237674713135,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2956790626049042,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384363174438,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.06159798800945282,],\n",
+      "[0.4502978026866913,],\n",
+      "[-0.46152129769325256,],\n",
+      "[0.9432777166366577,],\n",
+      "[0.8153313398361206,],\n",
+      "[-0.9385606050491333,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[5.258697456156369e-06,],\n",
+      "[-3.031771257155924e-06,],\n",
+      "[-6.504926568595693e-05,],\n",
+      "[2.1975089111947455e-05,],\n",
+      "[-0.4624532461166382,],\n",
+      "[-0.050088196992874146,],\n",
+      "[-0.5990619659423828,],\n",
+      "[-0.3156362771987915,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491656044498086,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27905237674713135,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0875276327133179,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384959220886,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.061135534197092056,],\n",
+      "[0.450347900390625,],\n",
+      "[-0.4609222412109375,],\n",
+      "[0.9435933232307434,],\n",
+      "[0.8154209852218628,],\n",
+      "[-0.9385614395141602,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27796486020088196,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5486401915550232,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391384959220886,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.061135534197092056,],\n",
+      "[0.450347900390625,],\n",
+      "[-0.4609222412109375,],\n",
+      "[0.9435933232307434,],\n",
+      "[0.8154209852218628,],\n",
+      "[-0.9385614395141602,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[5.280905497784261e-06,],\n",
+      "[-3.1324525480158627e-06,],\n",
+      "[-7.109055150067434e-05,],\n",
+      "[2.4560222300351597e-05,],\n",
+      "[-0.4554843306541443,],\n",
+      "[-0.047747138887643814,],\n",
+      "[-0.5369174480438232,],\n",
+      "[-0.3145487904548645,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491656626574695,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27796486020088196,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.48140287399292,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391385555267334,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.060680050402879715,],\n",
+      "[0.4503956437110901,],\n",
+      "[-0.4603853225708008,],\n",
+      "[0.9439078569412231,],\n",
+      "[0.815510630607605,],\n",
+      "[-0.938562273979187,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2764834463596344,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4006473422050476,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391385555267334,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.060680050402879715,],\n",
+      "[0.4503956437110901,],\n",
+      "[-0.4603853225708008,],\n",
+      "[0.9439078569412231,],\n",
+      "[0.815510630607605,],\n",
+      "[-0.938562273979187,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.3269266168645117e-06,],\n",
+      "[-2.0329002836660948e-06,],\n",
+      "[-4.87568904645741e-05,],\n",
+      "[1.722011438687332e-05,],\n",
+      "[-0.462004154920578,],\n",
+      "[-0.04930117353796959,],\n",
+      "[-0.5661994218826294,],\n",
+      "[-0.3130703866481781,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657208651304,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2764834463596344,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.265934705734253,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.060218047350645065,],\n",
+      "[0.45044493675231934,],\n",
+      "[-0.45981913805007935,],\n",
+      "[0.9442209005355835,],\n",
+      "[0.8156002759933472,],\n",
+      "[-0.9385631084442139,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2752175033092499,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08080232888460159,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.060218047350645065,],\n",
+      "[0.45044493675231934,],\n",
+      "[-0.45981913805007935,],\n",
+      "[0.9442209005355835,],\n",
+      "[0.8156002759933472,],\n",
+      "[-0.9385631084442139,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.1014661822628113e-06,],\n",
+      "[-6.928416382834257e-07,],\n",
+      "[-1.7569873307365924e-05,],\n",
+      "[6.344997927953955e-06,],\n",
+      "[-0.4768919348716736,],\n",
+      "[-0.05354950204491615,],\n",
+      "[-0.6729353070259094,],\n",
+      "[-0.31180745363235474,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657790727913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2752175033092499,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5685150027275085,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.05974115431308746,],\n",
+      "[0.45049849152565,],\n",
+      "[-0.4591462016105652,],\n",
+      "[0.9445326924324036,],\n",
+      "[0.8156899213790894,],\n",
+      "[-0.9385639429092407,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2746489942073822,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.00901651382446289,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.05974115431308746,],\n",
+      "[0.45049849152565,],\n",
+      "[-0.4591462016105652,],\n",
+      "[0.9445326924324036,],\n",
+      "[0.8156899213790894,],\n",
+      "[-0.9385639429092407,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.712539242111234e-07,],\n",
+      "[1.752526372911234e-07,],\n",
+      "[4.7066314436960965e-06,],\n",
+      "[-1.7390131006322918e-06,],\n",
+      "[-0.49099522829055786,],\n",
+      "[-0.05784092843532562,],\n",
+      "[-0.7920997738838196,],\n",
+      "[-0.31124192476272583,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657790727913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2746489942073822,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.18991079926490784,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.059250157326459885,],\n",
+      "[0.45055633783340454,],\n",
+      "[-0.458354115486145,],\n",
+      "[0.9448439478874207,],\n",
+      "[0.8157795667648315,],\n",
+      "[-0.9385647773742676,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2748388946056366,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07597973197698593,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.059250157326459885,],\n",
+      "[0.45055633783340454,],\n",
+      "[-0.458354115486145,],\n",
+      "[0.9448439478874207,],\n",
+      "[0.8157795667648315,],\n",
+      "[-0.9385647773742676,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.805014779980411e-07,],\n",
+      "[3.8430411564149836e-07,],\n",
+      "[1.0949157513095997e-05,],\n",
+      "[-4.141878889640793e-06,],\n",
+      "[-0.4968108832836151,],\n",
+      "[-0.059740833938121796,],\n",
+      "[-0.8477023243904114,],\n",
+      "[-0.311434805393219,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657790727913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2748388946056366,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5512891411781311,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.05875334516167641,],\n",
+      "[0.45061609148979187,],\n",
+      "[-0.4575064182281494,],\n",
+      "[0.9451553821563721,],\n",
+      "[0.8158692121505737,],\n",
+      "[-0.9385656118392944,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2753901779651642,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.022419605404138565,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.05875334516167641,],\n",
+      "[0.45061609148979187,],\n",
+      "[-0.4575064182281494,],\n",
+      "[0.9451553821563721,],\n",
+      "[0.8158692121505737,],\n",
+      "[-0.9385656118392944,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.3246855107572628e-07,],\n",
+      "[1.5751494686355727e-07,],\n",
+      "[4.764815457747318e-06,],\n",
+      "[-1.8459696775607881e-06,],\n",
+      "[-0.4923451244831085,],\n",
+      "[-0.05839259549975395,],\n",
+      "[-0.807870090007782,],\n",
+      "[-0.3119831085205078,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657790727913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2753901779651642,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2994636297225952,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.058260999619960785,],\n",
+      "[0.45067447423934937,],\n",
+      "[-0.45669853687286377,],\n",
+      "[0.9454673528671265,],\n",
+      "[0.8159588575363159,],\n",
+      "[-0.9385664463043213,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.27568963170051575,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.041900161653757095,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.058260999619960785,],\n",
+      "[0.45067447423934937,],\n",
+      "[-0.45669853687286377,],\n",
+      "[0.9454673528671265,],\n",
+      "[0.8159588575363159,],\n",
+      "[-0.9385664463043213,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.342912495123528e-07,],\n",
+      "[-1.62608998266478e-07,],\n",
+      "[-5.21960282640066e-06,],\n",
+      "[2.0705301722045988e-06,],\n",
+      "[-0.4816471040248871,],\n",
+      "[-0.05489422008395195,],\n",
+      "[-0.7025344371795654,],\n",
+      "[-0.3122825622558594,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657790727913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.27568963170051575,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4093908369541168,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.057779353111982346,],\n",
+      "[0.4507293701171875,],\n",
+      "[-0.45599600672721863,],\n",
+      "[0.9457796216011047,],\n",
+      "[0.8160485029220581,],\n",
+      "[-0.9385672807693481,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.275280237197876,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.32609570026397705,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.057779353111982346,],\n",
+      "[0.4507293701171875,],\n",
+      "[-0.45599600672721863,],\n",
+      "[0.9457796216011047,],\n",
+      "[0.8160485029220581,],\n",
+      "[-0.9385672807693481,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[4.818560341846023e-07,],\n",
+      "[-3.4333146459175623e-07,],\n",
+      "[-1.1675449059111997e-05,],\n",
+      "[4.739212272397708e-06,],\n",
+      "[-0.47193965315818787,],\n",
+      "[-0.051527950912714005,],\n",
+      "[-0.5948068499565125,],\n",
+      "[-0.31187018752098083,],\n",
+      "[-0.08964714407920837,],\n",
+      "[0.0008491657790727913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.275280237197876,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.1420964002609253,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.0573074147105217,],\n",
+      "[0.4507808983325958,],\n",
+      "[-0.455401211977005,],\n",
+      "[0.9460914731025696,],\n",
+      "[0.8161381483078003,],\n",
+      "[-0.938568115234375,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2741381525993347,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [3/10], Loss: 0.3261\n",
+      "loss_antes tensor([0.8737258315086365,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6537677645683289,],\n",
+      "[-0.13144570589065552,],\n",
+      "[0.5391386151313782,],\n",
+      "[-0.9985430240631104,],\n",
+      "[-0.0573074147105217,],\n",
+      "[0.4507808983325958,],\n",
+      "[-0.455401211977005,],\n",
+      "[0.9460914731025696,],\n",
+      "[0.8161381483078003,],\n",
+      "[-0.938568115234375,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3850361108779907,],\n",
+      "[-0.14731499552726746,],\n",
+      "[-0.4865354597568512,],\n",
+      "[0.08211658149957657,],\n",
+      "[-0.25991061329841614,],\n",
+      "[-0.03153153508901596,],\n",
+      "[-0.42358845472335815,],\n",
+      "[-0.19370557367801666,],\n",
+      "[-0.05401855707168579,],\n",
+      "[0.0005283018690533936,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2741381525993347,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8694645166397095,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6533827185630798,],\n",
+      "[-0.13129839301109314,],\n",
+      "[0.5396251678466797,],\n",
+      "[-0.9986251592636108,],\n",
+      "[-0.05704750493168831,],\n",
+      "[0.450812429189682,],\n",
+      "[-0.4549776315689087,],\n",
+      "[0.9462851881980896,],\n",
+      "[0.8161921501159668,],\n",
+      "[-0.938568651676178,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.272268682718277,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8657096028327942,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6533827185630798,],\n",
+      "[-0.13129839301109314,],\n",
+      "[0.5396251678466797,],\n",
+      "[-0.9986251592636108,],\n",
+      "[-0.05704750493168831,],\n",
+      "[0.450812429189682,],\n",
+      "[-0.4549776315689087,],\n",
+      "[0.9462851881980896,],\n",
+      "[0.8161921501159668,],\n",
+      "[-0.938568651676178,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.41247332096099854,],\n",
+      "[-0.14686597883701324,],\n",
+      "[-0.452229380607605,],\n",
+      "[0.0738791897892952,],\n",
+      "[-0.2611841857433319,],\n",
+      "[-0.030725639313459396,],\n",
+      "[-0.42261260747909546,],\n",
+      "[-0.20554061233997345,],\n",
+      "[-0.049619726836681366,],\n",
+      "[0.0009481781162321568,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.272268682718277,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8608695268630981,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6529702544212341,],\n",
+      "[-0.13115152716636658,],\n",
+      "[0.5400773882865906,],\n",
+      "[-0.9986990094184875,],\n",
+      "[-0.056786321103572845,],\n",
+      "[0.4508431553840637,],\n",
+      "[-0.4545550048351288,],\n",
+      "[0.9464907050132751,],\n",
+      "[0.8162417411804199,],\n",
+      "[-0.9385696053504944,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2704078257083893,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2942181825637817,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6529702544212341,],\n",
+      "[-0.13115152716636658,],\n",
+      "[0.5400773882865906,],\n",
+      "[-0.9986990094184875,],\n",
+      "[-0.056786321103572845,],\n",
+      "[0.4508431553840637,],\n",
+      "[-0.4545550048351288,],\n",
+      "[0.9464907050132751,],\n",
+      "[0.8162417411804199,],\n",
+      "[-0.9385696053504944,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5189365148544312,],\n",
+      "[-0.1730600893497467,],\n",
+      "[-0.5054467916488647,],\n",
+      "[0.08053752779960632,],\n",
+      "[-0.21210116147994995,],\n",
+      "[-0.02437910996377468,],\n",
+      "[-0.34446901082992554,],\n",
+      "[-0.19690236449241638,],\n",
+      "[-0.03792900592088699,],\n",
+      "[0.0012404497247189283,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2704078257083893,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.2752740383148193,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6524513363838196,],\n",
+      "[-0.13097846508026123,],\n",
+      "[0.5405828356742859,],\n",
+      "[-0.9987795352935791,],\n",
+      "[-0.056574221700429916,],\n",
+      "[0.45086753368377686,],\n",
+      "[-0.4542105495929718,],\n",
+      "[0.9466875791549683,],\n",
+      "[0.8162796497344971,],\n",
+      "[-0.9385708570480347,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2681325376033783,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.7930351495742798,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6524513363838196,],\n",
+      "[-0.13097846508026123,],\n",
+      "[0.5405828356742859,],\n",
+      "[-0.9987795352935791,],\n",
+      "[-0.056574221700429916,],\n",
+      "[0.45086753368377686,],\n",
+      "[-0.4542105495929718,],\n",
+      "[0.9466875791549683,],\n",
+      "[0.8162796497344971,],\n",
+      "[-0.9385708570480347,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5998929738998413,],\n",
+      "[-0.18932123482227325,],\n",
+      "[-0.5333840250968933,],\n",
+      "[0.08345745503902435,],\n",
+      "[-0.1658278852701187,],\n",
+      "[-0.018284350633621216,],\n",
+      "[-0.26787590980529785,],\n",
+      "[-0.19380617141723633,],\n",
+      "[-0.028788110241293907,],\n",
+      "[0.0015567566733807325,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2681325376033783,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.6780848503112793,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6518514156341553,],\n",
+      "[-0.1307891458272934,],\n",
+      "[0.5411162376403809,],\n",
+      "[-0.9988629817962646,],\n",
+      "[-0.05640839412808418,],\n",
+      "[0.4508858323097229,],\n",
+      "[-0.4539426863193512,],\n",
+      "[0.9468814134597778,],\n",
+      "[0.8163084387779236,],\n",
+      "[-0.9385724067687988,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2654544413089752,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.7533721923828125,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6518514156341553,],\n",
+      "[-0.1307891458272934,],\n",
+      "[0.5411162376403809,],\n",
+      "[-0.9988629817962646,],\n",
+      "[-0.05640839412808418,],\n",
+      "[0.4508858323097229,],\n",
+      "[-0.4539426863193512,],\n",
+      "[0.9468814134597778,],\n",
+      "[0.8163084387779236,],\n",
+      "[-0.9385724067687988,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5563762187957764,],\n",
+      "[-0.1682703197002411,],\n",
+      "[-0.46462640166282654,],\n",
+      "[0.07185513526201248,],\n",
+      "[-0.17221079766750336,],\n",
+      "[-0.01789737120270729,],\n",
+      "[-0.27057647705078125,],\n",
+      "[-0.2099132239818573,],\n",
+      "[-0.03185604512691498,],\n",
+      "[0.0019519524648785591,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2654544413089752,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.648298978805542,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6512950658798218,],\n",
+      "[-0.13062088191509247,],\n",
+      "[0.541580855846405,],\n",
+      "[-0.9989348649978638,],\n",
+      "[-0.056236185133457184,],\n",
+      "[0.4509037435054779,],\n",
+      "[-0.4536721110343933,],\n",
+      "[0.9470913410186768,],\n",
+      "[0.8163402676582336,],\n",
+      "[-0.9385743737220764,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.26280614733695984,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.038375735282898,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6512950658798218,],\n",
+      "[-0.13062088191509247,],\n",
+      "[0.541580855846405,],\n",
+      "[-0.9989348649978638,],\n",
+      "[-0.056236185133457184,],\n",
+      "[0.4509037435054779,],\n",
+      "[-0.4536721110343933,],\n",
+      "[0.9470913410186768,],\n",
+      "[0.8163402676582336,],\n",
+      "[-0.9385743737220764,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3844916522502899,],\n",
+      "[-0.11285840719938278,],\n",
+      "[-0.3099977374076843,],\n",
+      "[0.04769527167081833,],\n",
+      "[-0.24647027254104614,],\n",
+      "[-0.024798844009637833,],\n",
+      "[-0.3807747960090637,],\n",
+      "[-0.24115899205207825,],\n",
+      "[-0.04743883013725281,],\n",
+      "[0.0023490716703236103,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.26280614733695984,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.0380141735076904,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6509105563163757,],\n",
+      "[-0.13050802052021027,],\n",
+      "[0.5418908596038818,],\n",
+      "[-0.9989825487136841,],\n",
+      "[-0.055989716202020645,],\n",
+      "[0.45092853903770447,],\n",
+      "[-0.4532913267612457,],\n",
+      "[0.947332501411438,],\n",
+      "[0.8163877129554749,],\n",
+      "[-0.9385766983032227,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2607681453227997,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.28396037220954895,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6509105563163757,],\n",
+      "[-0.13050802052021027,],\n",
+      "[0.5418908596038818,],\n",
+      "[-0.9989825487136841,],\n",
+      "[-0.055989716202020645,],\n",
+      "[0.45092853903770447,],\n",
+      "[-0.4532913267612457,],\n",
+      "[0.947332501411438,],\n",
+      "[0.8163877129554749,],\n",
+      "[-0.9385766983032227,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.1736600250005722,],\n",
+      "[-0.050011392682790756,],\n",
+      "[-0.1385928839445114,],\n",
+      "[0.021347150206565857,],\n",
+      "[-0.3603517711162567,],\n",
+      "[-0.03603281453251839,],\n",
+      "[-0.5589380264282227,],\n",
+      "[-0.27360111474990845,],\n",
+      "[-0.06685624271631241,],\n",
+      "[0.0026367735117673874,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2607681453227997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.065759301185608,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6507368683815002,],\n",
+      "[-0.1304580122232437,],\n",
+      "[0.5420294404029846,],\n",
+      "[-0.9990038871765137,],\n",
+      "[-0.055629365146160126,],\n",
+      "[0.4509645700454712,],\n",
+      "[-0.4527323842048645,],\n",
+      "[0.947606086730957,],\n",
+      "[0.8164545893669128,],\n",
+      "[-0.9385793209075928,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2597023844718933,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.008672062307596207,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6507368683815002,],\n",
+      "[-0.1304580122232437,],\n",
+      "[0.5420294404029846,],\n",
+      "[-0.9990038871765137,],\n",
+      "[-0.055629365146160126,],\n",
+      "[0.4509645700454712,],\n",
+      "[-0.4527323842048645,],\n",
+      "[0.947606086730957,],\n",
+      "[0.8164545893669128,],\n",
+      "[-0.9385793209075928,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.02535470947623253,],\n",
+      "[-0.007224178873002529,],\n",
+      "[-0.020461831241846085,],\n",
+      "[0.0031723349820822477,],\n",
+      "[-0.46049612760543823,],\n",
+      "[-0.04623739421367645,],\n",
+      "[-0.7212854027748108,],\n",
+      "[-0.2950546443462372,],\n",
+      "[-0.0804973617196083,],\n",
+      "[0.002774594584479928,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2597023844718933,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.1862480193376541,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6507115364074707,],\n",
+      "[-0.1304507851600647,],\n",
+      "[0.5420498847961426,],\n",
+      "[-0.9990070462226868,],\n",
+      "[-0.05516887083649635,],\n",
+      "[0.4510107934474945,],\n",
+      "[-0.4520111083984375,],\n",
+      "[0.9479011297225952,],\n",
+      "[0.8165351152420044,],\n",
+      "[-0.9385821223258972,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2595161497592926,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.008510066196322441,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6507115364074707,],\n",
+      "[-0.1304507851600647,],\n",
+      "[0.5420498847961426,],\n",
+      "[-0.9990070462226868,],\n",
+      "[-0.05516887083649635,],\n",
+      "[0.4510107934474945,],\n",
+      "[-0.4520111083984375,],\n",
+      "[0.9479011297225952,],\n",
+      "[0.8165351152420044,],\n",
+      "[-0.9385821223258972,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0204161349684,],\n",
+      "[0.005793711636215448,],\n",
+      "[0.016957702115178108,],\n",
+      "[-0.0026575077790766954,],\n",
+      "[-0.501480221748352,],\n",
+      "[-0.05062917619943619,],\n",
+      "[-0.7902026772499084,],\n",
+      "[-0.3017156720161438,],\n",
+      "[-0.08467802405357361,],\n",
+      "[0.002802836010232568,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2595161497592926,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.18450018763542175,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6507319808006287,],\n",
+      "[-0.1304565817117691,],\n",
+      "[0.5420328974723816,],\n",
+      "[-0.9990043640136719,],\n",
+      "[-0.0546673908829689,],\n",
+      "[0.4510614275932312,],\n",
+      "[-0.45122089982032776,],\n",
+      "[0.9482028484344482,],\n",
+      "[0.8166198134422302,],\n",
+      "[-0.9385849237442017,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2597006559371948,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0015319677768275142,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6507319808006287,],\n",
+      "[-0.1304565817117691,],\n",
+      "[0.5420328974723816,],\n",
+      "[-0.9990043640136719,],\n",
+      "[-0.0546673908829689,],\n",
+      "[0.4510614275932312,],\n",
+      "[-0.45122089982032776,],\n",
+      "[0.9482028484344482,],\n",
+      "[0.8166198134422302,],\n",
+      "[-0.9385849237442017,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0068845683708786964,],\n",
+      "[-0.001958406064659357,],\n",
+      "[-0.0059710354544222355,],\n",
+      "[0.0009488008799962699,],\n",
+      "[-0.47292008996009827,],\n",
+      "[-0.047541145235300064,],\n",
+      "[-0.7412534356117249,],\n",
+      "[-0.2977646291255951,],\n",
+      "[-0.08221732079982758,],\n",
+      "[0.0027871050406247377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2597006559371948,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.0782807394862175,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6507250666618347,],\n",
+      "[-0.1304546296596527,],\n",
+      "[0.5420388579368591,],\n",
+      "[-0.9990053176879883,],\n",
+      "[-0.05419446900486946,],\n",
+      "[0.4511089622974396,],\n",
+      "[-0.450479656457901,],\n",
+      "[0.9485006332397461,],\n",
+      "[0.8167020082473755,],\n",
+      "[-0.9385877251625061,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25962236523628235,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13371635973453522,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6507250666618347,],\n",
+      "[-0.1304546296596527,],\n",
+      "[0.5420388579368591,],\n",
+      "[-0.9990053176879883,],\n",
+      "[-0.05419446900486946,],\n",
+      "[0.4511089622974396,],\n",
+      "[-0.450479656457901,],\n",
+      "[0.9485006332397461,],\n",
+      "[0.8167020082473755,],\n",
+      "[-0.9385877251625061,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.05018254742026329,],\n",
+      "[-0.014410608448088169,],\n",
+      "[-0.0459856316447258,],\n",
+      "[0.007426028605550528,],\n",
+      "[-0.40469419956207275,],\n",
+      "[-0.039616119116544724,],\n",
+      "[-0.6187307238578796,],\n",
+      "[-0.2907043993473053,],\n",
+      "[-0.07838679105043411,],\n",
+      "[0.002772890729829669,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25962236523628235,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7313445806503296,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6506748795509338,],\n",
+      "[-0.13044022023677826,],\n",
+      "[0.5420848727226257,],\n",
+      "[-0.9990127682685852,],\n",
+      "[-0.05378977581858635,],\n",
+      "[0.4511485695838928,],\n",
+      "[-0.44986093044281006,],\n",
+      "[0.9487913250923157,],\n",
+      "[0.8167803883552551,],\n",
+      "[-0.9385905265808105,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2588910162448883,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.4286006689071655,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6506748795509338,],\n",
+      "[-0.13044022023677826,],\n",
+      "[0.5420848727226257,],\n",
+      "[-0.9990127682685852,],\n",
+      "[-0.05378977581858635,],\n",
+      "[0.4511485695838928,],\n",
+      "[-0.44986093044281006,],\n",
+      "[0.9487913250923157,],\n",
+      "[0.8167803883552551,],\n",
+      "[-0.9385905265808105,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.06905116885900497,],\n",
+      "[-0.020160727202892303,],\n",
+      "[-0.06748296320438385,],\n",
+      "[0.011094876565039158,],\n",
+      "[-0.34758538007736206,],\n",
+      "[-0.03245600685477257,],\n",
+      "[-0.5096127986907959,],\n",
+      "[-0.2864218056201935,],\n",
+      "[-0.07677358388900757,],\n",
+      "[0.002775121247395873,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2588910162448883,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.3093516826629639,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6506058573722839,],\n",
+      "[-0.13042005896568298,],\n",
+      "[0.5421523451805115,],\n",
+      "[-0.9990238547325134,],\n",
+      "[-0.05344219133257866,],\n",
+      "[0.451181024312973,],\n",
+      "[-0.44935131072998047,],\n",
+      "[0.9490777254104614,],\n",
+      "[0.8168571591377258,],\n",
+      "[-0.938593327999115,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2575816512107849,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.49372512102127075,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6506058573722839,],\n",
+      "[-0.13042005896568298,],\n",
+      "[0.5421523451805115,],\n",
+      "[-0.9990238547325134,],\n",
+      "[-0.05344219133257866,],\n",
+      "[0.451181024312973,],\n",
+      "[-0.44935131072998047,],\n",
+      "[0.9490777254104614,],\n",
+      "[0.8168571591377258,],\n",
+      "[-0.938593327999115,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.056300677359104156,],\n",
+      "[-0.016808761283755302,],\n",
+      "[-0.05908764898777008,],\n",
+      "[0.009905463084578514,],\n",
+      "[-0.3416728377342224,],\n",
+      "[-0.031002119183540344,],\n",
+      "[-0.4907863140106201,],\n",
+      "[-0.28675904870033264,],\n",
+      "[-0.07789838314056396,],\n",
+      "[0.0027889152988791466,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2575816512107849,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4053114652633667,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505495309829712,],\n",
+      "[-0.13040325045585632,],\n",
+      "[0.5422114133834839,],\n",
+      "[-0.9990337491035461,],\n",
+      "[-0.05310051888227463,],\n",
+      "[0.4512120187282562,],\n",
+      "[-0.4488605260848999,],\n",
+      "[0.9493644833564758,],\n",
+      "[0.8169350624084473,],\n",
+      "[-0.9385961294174194,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2561763525009155,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2096036672592163,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505495309829712,],\n",
+      "[-0.13040325045585632,],\n",
+      "[0.5422114133834839,],\n",
+      "[-0.9990337491035461,],\n",
+      "[-0.05310051888227463,],\n",
+      "[0.4512120187282562,],\n",
+      "[-0.4488605260848999,],\n",
+      "[0.9493644833564758,],\n",
+      "[0.8169350624084473,],\n",
+      "[-0.9385961294174194,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.02763313800096512,],\n",
+      "[-0.008460969664156437,],\n",
+      "[-0.031288862228393555,],\n",
+      "[0.005355154629796743,],\n",
+      "[-0.3926181495189667,],\n",
+      "[-0.03633623942732811,],\n",
+      "[-0.582098662853241,],\n",
+      "[-0.290177583694458,],\n",
+      "[-0.08032478392124176,],\n",
+      "[0.002803932409733534,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2561763525009155,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9156509041786194,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505218744277954,],\n",
+      "[-0.13039478659629822,],\n",
+      "[0.542242705821991,],\n",
+      "[-0.9990391135215759,],\n",
+      "[-0.05270789936184883,],\n",
+      "[0.4512483477592468,],\n",
+      "[-0.44827842712402344,],\n",
+      "[0.9496546387672424,],\n",
+      "[0.8170154094696045,],\n",
+      "[-0.9385989308357239,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.255260705947876,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0025654027704149485,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505218744277954,],\n",
+      "[-0.13039478659629822,],\n",
+      "[0.542242705821991,],\n",
+      "[-0.9990391135215759,],\n",
+      "[-0.05270789936184883,],\n",
+      "[0.4512483477592468,],\n",
+      "[-0.44827842712402344,],\n",
+      "[0.9496546387672424,],\n",
+      "[0.8170154094696045,],\n",
+      "[-0.9385989308357239,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.002289539435878396,],\n",
+      "[-0.0007190198521129787,],\n",
+      "[-0.0028047964442521334,],\n",
+      "[0.0004906325484625995,],\n",
+      "[-0.4706333875656128,],\n",
+      "[-0.04533083736896515,],\n",
+      "[-0.7347719669342041,],\n",
+      "[-0.29404836893081665,],\n",
+      "[-0.08241638541221619,],\n",
+      "[0.0028134463354945183,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.255260705947876,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.10129959881305695,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.650519609451294,],\n",
+      "[-0.1303940713405609,],\n",
+      "[0.5422455072402954,],\n",
+      "[-0.9990395903587341,],\n",
+      "[-0.052237264811992645,],\n",
+      "[0.4512936770915985,],\n",
+      "[-0.4475436508655548,],\n",
+      "[0.9499486684799194,],\n",
+      "[0.8170978426933289,],\n",
+      "[-0.9386017322540283,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2551594078540802,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08187684416770935,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.650519609451294,],\n",
+      "[-0.1303940713405609,],\n",
+      "[0.5422455072402954,],\n",
+      "[-0.9990395903587341,],\n",
+      "[-0.052237264811992645,],\n",
+      "[0.4512936770915985,],\n",
+      "[-0.4475436508655548,],\n",
+      "[0.9499486684799194,],\n",
+      "[0.8170978426933289,],\n",
+      "[-0.9386017322540283,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.009648008272051811,],\n",
+      "[0.0031028103549033403,],\n",
+      "[0.012808453291654587,],\n",
+      "[-0.0022919941693544388,],\n",
+      "[-0.5310860276222229,],\n",
+      "[-0.052842870354652405,],\n",
+      "[-0.8615264296531677,],\n",
+      "[-0.29650571942329407,],\n",
+      "[-0.08337593078613281,],\n",
+      "[0.0028163946699351072,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2551594078540802,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5722822546958923,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505292654037476,],\n",
+      "[-0.13039717078208923,],\n",
+      "[0.5422326922416687,],\n",
+      "[-0.9990373253822327,],\n",
+      "[-0.05170617997646332,],\n",
+      "[0.4513465166091919,],\n",
+      "[-0.4466821253299713,],\n",
+      "[0.950245201587677,],\n",
+      "[0.8171812295913696,],\n",
+      "[-0.9386045336723328,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2557317018508911,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.12928272783756256,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505292654037476,],\n",
+      "[-0.13039717078208923,],\n",
+      "[0.5422326922416687,],\n",
+      "[-0.9990373253822327,],\n",
+      "[-0.05170617997646332,],\n",
+      "[0.4513465166091919,],\n",
+      "[-0.4466821253299713,],\n",
+      "[0.950245201587677,],\n",
+      "[0.8171812295913696,],\n",
+      "[-0.9386045336723328,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.009015426971018314,],\n",
+      "[0.002964958781376481,],\n",
+      "[0.012984334491193295,],\n",
+      "[-0.0023785140365362167,],\n",
+      "[-0.5421279668807983,],\n",
+      "[-0.05467035621404648,],\n",
+      "[-0.8895916938781738,],\n",
+      "[-0.2970886826515198,],\n",
+      "[-0.08331989496946335,],\n",
+      "[0.0028153134044259787,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2557317018508911,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.719118058681488,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505382657051086,],\n",
+      "[-0.1304001361131668,],\n",
+      "[0.5422196984291077,],\n",
+      "[-0.9990349411964417,],\n",
+      "[-0.05116405338048935,],\n",
+      "[0.4514011740684509,],\n",
+      "[-0.4457925260066986,],\n",
+      "[0.9505422711372375,],\n",
+      "[0.8172645568847656,],\n",
+      "[-0.9386073350906372,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2564508318901062,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.018002483993768692,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505382657051086,],\n",
+      "[-0.1304001361131668,],\n",
+      "[0.5422196984291077,],\n",
+      "[-0.9990349411964417,],\n",
+      "[-0.05116405338048935,],\n",
+      "[0.4514011740684509,],\n",
+      "[-0.4457925260066986,],\n",
+      "[0.9505422711372375,],\n",
+      "[0.8172645568847656,],\n",
+      "[-0.9386073350906372,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.002495324471965432,],\n",
+      "[0.0008393487078137696,],\n",
+      "[0.0039025719743222,],\n",
+      "[-0.0007321533048525453,],\n",
+      "[-0.5026026964187622,],\n",
+      "[-0.04991654306650162,],\n",
+      "[-0.8050212860107422,],\n",
+      "[-0.2963193953037262,],\n",
+      "[-0.08281224966049194,],\n",
+      "[0.002813245402649045,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2564508318901062,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.26834636926651,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505407691001892,],\n",
+      "[-0.13040097057819366,],\n",
+      "[0.5422158241271973,],\n",
+      "[-0.9990342259407043,],\n",
+      "[-0.05066145211458206,],\n",
+      "[0.4514510929584503,],\n",
+      "[-0.4449875056743622,],\n",
+      "[0.950838565826416,],\n",
+      "[0.8173473477363586,],\n",
+      "[-0.9386101365089417,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.256719172000885,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.060415755957365036,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505407691001892,],\n",
+      "[-0.13040097057819366,],\n",
+      "[0.5422158241271973,],\n",
+      "[-0.9990342259407043,],\n",
+      "[-0.05066145211458206,],\n",
+      "[0.4514510929584503,],\n",
+      "[-0.4449875056743622,],\n",
+      "[0.950838565826416,],\n",
+      "[0.8173473477363586,],\n",
+      "[-0.9386101365089417,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.003383215982466936,],\n",
+      "[-0.0011662100441753864,],\n",
+      "[-0.005751276854425669,],\n",
+      "[0.0011052159825339913,],\n",
+      "[-0.44101670384407043,],\n",
+      "[-0.04167936369776726,],\n",
+      "[-0.6618363261222839,],\n",
+      "[-0.2950299084186554,],\n",
+      "[-0.08236756175756454,],\n",
+      "[0.0028119792696088552,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.256719172000885,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4915923476219177,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.650537371635437,],\n",
+      "[-0.13039980828762054,],\n",
+      "[0.5422215461730957,],\n",
+      "[-0.9990353584289551,],\n",
+      "[-0.05022043362259865,],\n",
+      "[0.4514927864074707,],\n",
+      "[-0.44432565569877625,],\n",
+      "[0.9511336088180542,],\n",
+      "[0.8174297213554382,],\n",
+      "[-0.9386129379272461,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2562275826931,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2955131232738495,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.650537371635437,],\n",
+      "[-0.13039980828762054,],\n",
+      "[0.5422215461730957,],\n",
+      "[-0.9990353584289551,],\n",
+      "[-0.05022043362259865,],\n",
+      "[0.4514927864074707,],\n",
+      "[-0.44432565569877625,],\n",
+      "[0.9511336088180542,],\n",
+      "[0.8174297213554382,],\n",
+      "[-0.9386129379272461,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00552814407274127,],\n",
+      "[-0.0019582563545554876,],\n",
+      "[-0.01022367924451828,],\n",
+      "[0.0020122851710766554,],\n",
+      "[-0.3969068229198456,],\n",
+      "[-0.035218607634305954,],\n",
+      "[-0.5494343042373657,],\n",
+      "[-0.2938351333141327,],\n",
+      "[-0.082210972905159,],\n",
+      "[0.002811869140714407,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2562275826931,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0872234106063843,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505318284034729,],\n",
+      "[-0.13039785623550415,],\n",
+      "[0.5422317981719971,],\n",
+      "[-0.9990373849868774,],\n",
+      "[-0.04982352629303932,],\n",
+      "[0.45152801275253296,],\n",
+      "[-0.4437762200832367,],\n",
+      "[0.9514274597167969,],\n",
+      "[0.8175119161605835,],\n",
+      "[-0.9386157393455505,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25514036417007446,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.32887646555900574,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505318284034729,],\n",
+      "[-0.13039785623550415,],\n",
+      "[0.5422317981719971,],\n",
+      "[-0.9990373849868774,],\n",
+      "[-0.04982352629303932,],\n",
+      "[0.45152801275253296,],\n",
+      "[-0.4437762200832367,],\n",
+      "[0.9514274597167969,],\n",
+      "[0.8175119161605835,],\n",
+      "[-0.9386157393455505,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.004303389228880405,],\n",
+      "[-0.0015698372153565288,],\n",
+      "[-0.008663441054522991,],\n",
+      "[0.0017463197000324726,],\n",
+      "[-0.3961089551448822,],\n",
+      "[-0.03450153023004532,],\n",
+      "[-0.5378785133361816,],\n",
+      "[-0.293015718460083,],\n",
+      "[-0.08230286091566086,],\n",
+      "[0.0028124526143074036,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25514036417007446,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.146955132484436,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505275368690491,],\n",
+      "[-0.1303962916135788,],\n",
+      "[0.5422404408454895,],\n",
+      "[-0.9990391135215759,],\n",
+      "[-0.04942741617560387,],\n",
+      "[0.4515625238418579,],\n",
+      "[-0.44323834776878357,],\n",
+      "[0.9517204761505127,],\n",
+      "[0.8175942301750183,],\n",
+      "[-0.938618540763855,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.253993421792984,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0980929359793663,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505275368690491,],\n",
+      "[-0.1303962916135788,],\n",
+      "[0.5422404408454895,],\n",
+      "[-0.9990391135215759,],\n",
+      "[-0.04942741617560387,],\n",
+      "[0.4515625238418579,],\n",
+      "[-0.44323834776878357,],\n",
+      "[0.9517204761505127,],\n",
+      "[0.8175942301750183,],\n",
+      "[-0.938618540763855,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0017329700058326125,],\n",
+      "[-0.0006511903484351933,],\n",
+      "[-0.003798276884481311,],\n",
+      "[0.0007841449696570635,],\n",
+      "[-0.43616268038749695,],\n",
+      "[-0.03972421586513519,],\n",
+      "[-0.6354001760482788,],\n",
+      "[-0.2926482856273651,],\n",
+      "[-0.0824875757098198,],\n",
+      "[0.0028131084982305765,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.253993421792984,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6263953447341919,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505258083343506,],\n",
+      "[-0.13039563596248627,],\n",
+      "[0.5422442555427551,],\n",
+      "[-0.999039888381958,],\n",
+      "[-0.04899125173687935,],\n",
+      "[0.4516022503376007,],\n",
+      "[-0.4426029622554779,],\n",
+      "[0.9520131349563599,],\n",
+      "[0.8176767230033875,],\n",
+      "[-0.9386213421821594,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25336703658103943,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.006798306480050087,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505258083343506,],\n",
+      "[-0.13039563596248627,],\n",
+      "[0.5422442555427551,],\n",
+      "[-0.999039888381958,],\n",
+      "[-0.04899125173687935,],\n",
+      "[0.4516022503376007,],\n",
+      "[-0.4426029622554779,],\n",
+      "[0.9520131349563599,],\n",
+      "[0.8176767230033875,],\n",
+      "[-0.9386213421821594,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0003362519491929561,],\n",
+      "[0.0001299469731748104,],\n",
+      "[0.0008021406829357147,],\n",
+      "[-0.00016966329712886363,],\n",
+      "[-0.49064117670059204,],\n",
+      "[-0.047525856643915176,],\n",
+      "[-0.7837758660316467,],\n",
+      "[-0.29274216294288635,],\n",
+      "[-0.08263211697340012,],\n",
+      "[0.0028134998865425587,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25336703658103943,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.1649036854505539,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505261659622192,],\n",
+      "[-0.130395770072937,],\n",
+      "[0.542243480682373,],\n",
+      "[-0.9990397095680237,],\n",
+      "[-0.04850061237812042,],\n",
+      "[0.4516497850418091,],\n",
+      "[-0.4418191909790039,],\n",
+      "[0.9523058533668518,],\n",
+      "[0.8177593350410461,],\n",
+      "[-0.9386241436004639,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25353193283081055,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1412171721458435,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505261659622192,],\n",
+      "[-0.130395770072937,],\n",
+      "[0.542243480682373,],\n",
+      "[-0.9990397095680237,],\n",
+      "[-0.04850061237812042,],\n",
+      "[0.4516497850418091,],\n",
+      "[-0.4418191909790039,],\n",
+      "[0.9523058533668518,],\n",
+      "[0.8177593350410461,],\n",
+      "[-0.9386241436004639,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.001129232463426888,],\n",
+      "[0.0004477282636798918,],\n",
+      "[0.0029304532799869776,],\n",
+      "[-0.0006353597273118794,],\n",
+      "[-0.5272943377494812,],\n",
+      "[-0.053254634141922,],\n",
+      "[-0.8938031196594238,],\n",
+      "[-0.2932344675064087,],\n",
+      "[-0.08268604427576065,],\n",
+      "[0.002813576953485608,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25353193283081055,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7515773177146912,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.65052729845047,],\n",
+      "[-0.13039621710777283,],\n",
+      "[0.542240560054779,],\n",
+      "[-0.9990390539169312,],\n",
+      "[-0.04797331988811493,],\n",
+      "[0.4517030417919159,],\n",
+      "[-0.44092538952827454,],\n",
+      "[0.9525991082191467,],\n",
+      "[0.8178420066833496,],\n",
+      "[-0.9386269450187683,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25428351759910583,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15262602269649506,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.65052729845047,],\n",
+      "[-0.13039621710777283,],\n",
+      "[0.542240560054779,],\n",
+      "[-0.9990390539169312,],\n",
+      "[-0.04797331988811493,],\n",
+      "[0.4517030417919159,],\n",
+      "[-0.44092538952827454,],\n",
+      "[0.9525991082191467,],\n",
+      "[0.8178420066833496,],\n",
+      "[-0.9386269450187683,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0008648190414533019,],\n",
+      "[0.00035116681829094887,],\n",
+      "[0.0024403929710388184,],\n",
+      "[-0.0005425880081020296,],\n",
+      "[-0.5270733833312988,],\n",
+      "[-0.053644485771656036,],\n",
+      "[-0.8994597792625427,],\n",
+      "[-0.2939070165157318,],\n",
+      "[-0.08266834914684296,],\n",
+      "[0.0028134803287684917,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25428351759910583,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7813469171524048,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505281925201416,],\n",
+      "[-0.13039657473564148,],\n",
+      "[0.5422381162643433,],\n",
+      "[-0.9990385174751282,],\n",
+      "[-0.047446247190237045,],\n",
+      "[0.45175668597221375,],\n",
+      "[-0.4400259256362915,],\n",
+      "[0.9528930187225342,],\n",
+      "[0.8179246783256531,],\n",
+      "[-0.9386297464370728,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25506487488746643,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.01431826688349247,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505281925201416,],\n",
+      "[-0.13039657473564148,],\n",
+      "[0.5422381162643433,],\n",
+      "[-0.9990385174751282,],\n",
+      "[-0.047446247190237045,],\n",
+      "[0.45175668597221375,],\n",
+      "[-0.4400259256362915,],\n",
+      "[0.9528930187225342,],\n",
+      "[0.8179246783256531,],\n",
+      "[-0.9386297464370728,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00019508400873746723,],\n",
+      "[8.112156501738355e-05,],\n",
+      "[0.0005985616007819772,],\n",
+      "[-0.00013648094318341464,],\n",
+      "[-0.4937626123428345,],\n",
+      "[-0.04864691570401192,],\n",
+      "[-0.7983520030975342,],\n",
+      "[-0.29440397024154663,],\n",
+      "[-0.08262360095977783,],\n",
+      "[0.0028133748564869165,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25506487488746643,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.23931823670864105,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505283713340759,],\n",
+      "[-0.13039664924144745,],\n",
+      "[0.5422375202178955,],\n",
+      "[-0.9990383982658386,],\n",
+      "[-0.04695248603820801,],\n",
+      "[0.45180532336235046,],\n",
+      "[-0.4392275810241699,],\n",
+      "[0.9531874060630798,],\n",
+      "[0.8180072903633118,],\n",
+      "[-0.9386325478553772,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2553041875362396,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07361412793397903,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505283713340759,],\n",
+      "[-0.13039664924144745,],\n",
+      "[0.5422375202178955,],\n",
+      "[-0.9990383982658386,],\n",
+      "[-0.04695248603820801,],\n",
+      "[0.45180532336235046,],\n",
+      "[-0.4392275810241699,],\n",
+      "[0.9531874060630798,],\n",
+      "[0.8180072903633118,],\n",
+      "[-0.9386325478553772,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00032570306211709976,],\n",
+      "[-0.0001389286044286564,],\n",
+      "[-0.0010869338875636458,],\n",
+      "[0.0002540796995162964,],\n",
+      "[-0.4502190947532654,],\n",
+      "[-0.041486941277980804,],\n",
+      "[-0.6530008316040039,],\n",
+      "[-0.294383704662323,],\n",
+      "[-0.08259022980928421,],\n",
+      "[0.002813320839777589,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2553041875362396,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5426381230354309,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.650528073310852,],\n",
+      "[-0.1303965151309967,],\n",
+      "[0.5422385931015015,],\n",
+      "[-0.9990386366844177,],\n",
+      "[-0.04650226607918739,],\n",
+      "[0.45184680819511414,],\n",
+      "[-0.43857458233833313,],\n",
+      "[0.9534817934036255,],\n",
+      "[0.8180899024009705,],\n",
+      "[-0.9386353492736816,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25476154685020447,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.29552310705184937,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.650528073310852,],\n",
+      "[-0.1303965151309967,],\n",
+      "[0.5422385931015015,],\n",
+      "[-0.9990386366844177,],\n",
+      "[-0.04650226607918739,],\n",
+      "[0.45184680819511414,],\n",
+      "[-0.43857458233833313,],\n",
+      "[0.9534817934036255,],\n",
+      "[0.8180899024009705,],\n",
+      "[-0.9386353492736816,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00048042205162346363,],\n",
+      "[-0.00021071212540846318,],\n",
+      "[-0.0017446931451559067,],\n",
+      "[0.0004178822273388505,],\n",
+      "[-0.4234596788883209,],\n",
+      "[-0.03663157299160957,],\n",
+      "[-0.5519300103187561,],\n",
+      "[-0.2937432825565338,],\n",
+      "[-0.08258029073476791,],\n",
+      "[0.002813309198245406,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25476154685020447,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.087241530418396,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505275964736938,],\n",
+      "[-0.13039630651474,],\n",
+      "[0.5422403216362,],\n",
+      "[-0.9990390539169312,],\n",
+      "[-0.04607880488038063,],\n",
+      "[0.4518834352493286,],\n",
+      "[-0.438022643327713,],\n",
+      "[0.9537755250930786,],\n",
+      "[0.8181724548339844,],\n",
+      "[-0.9386381506919861,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25367429852485657,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.28175926208496094,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505275964736938,],\n",
+      "[-0.13039630651474,],\n",
+      "[0.5422403216362,],\n",
+      "[-0.9990390539169312,],\n",
+      "[-0.04607880488038063,],\n",
+      "[0.4518834352493286,],\n",
+      "[-0.438022643327713,],\n",
+      "[0.9537755250930786,],\n",
+      "[0.8181724548339844,],\n",
+      "[-0.9386381506919861,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00034530434641055763,],\n",
+      "[-0.00015598503523506224,],\n",
+      "[-0.0013651176122948527,],\n",
+      "[0.0003348842146806419,],\n",
+      "[-0.4278196692466736,],\n",
+      "[-0.0368928499519825,],\n",
+      "[-0.5562716126441956,],\n",
+      "[-0.29271653294563293,],\n",
+      "[-0.08258886635303497,],\n",
+      "[0.0028133485466241837,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25367429852485657,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0616194009780884,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04565098509192467,],\n",
+      "[0.4519203305244446,],\n",
+      "[-0.4374663829803467,],\n",
+      "[0.9540682435035706,],\n",
+      "[0.8182550668716431,],\n",
+      "[-0.9386409521102905,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2526126801967621,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.05750023201107979,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04565098509192467,],\n",
+      "[0.4519203305244446,],\n",
+      "[-0.4374663829803467,],\n",
+      "[0.9540682435035706,],\n",
+      "[0.8182550668716431,],\n",
+      "[-0.9386409521102905,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00011481753608677536,],\n",
+      "[-5.341498035704717e-05,],\n",
+      "[-0.0004941108636558056,],\n",
+      "[0.0001241487480001524,],\n",
+      "[-0.4575651288032532,],\n",
+      "[-0.041782986372709274,],\n",
+      "[-0.6635643839836121,],\n",
+      "[-0.29178673028945923,],\n",
+      "[-0.08260350674390793,],\n",
+      "[0.002813382539898157,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2526126801967621,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.47958433628082275,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505271196365356,],\n",
+      "[-0.13039609789848328,],\n",
+      "[0.542242169380188,],\n",
+      "[-0.9990395307540894,],\n",
+      "[-0.045193418860435486,],\n",
+      "[0.45196211338043213,],\n",
+      "[-0.43680280447006226,],\n",
+      "[0.9543600082397461,],\n",
+      "[0.8183376789093018,],\n",
+      "[-0.938643753528595,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2521331012248993,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.023498618975281715,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505271196365356,],\n",
+      "[-0.13039609789848328,],\n",
+      "[0.542242169380188,],\n",
+      "[-0.9990395307540894,],\n",
+      "[-0.045193418860435486,],\n",
+      "[0.45196211338043213,],\n",
+      "[-0.43680280447006226,],\n",
+      "[0.9543600082397461,],\n",
+      "[0.8183376789093018,],\n",
+      "[-0.938643753528595,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.4025174904381856e-05,],\n",
+      "[2.5837694920483045e-05,],\n",
+      "[0.00025296147214248776,],\n",
+      "[-6.512566324090585e-05,],\n",
+      "[-0.49293357133865356,],\n",
+      "[-0.04812702536582947,],\n",
+      "[-0.8087178468704224,],\n",
+      "[-0.2914206385612488,],\n",
+      "[-0.08261311054229736,],\n",
+      "[0.0028133857995271683,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2521331012248993,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3065853714942932,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505271792411804,],\n",
+      "[-0.13039612770080566,],\n",
+      "[0.5422419309616089,],\n",
+      "[-0.9990394711494446,],\n",
+      "[-0.04470048472285271,],\n",
+      "[0.45201024413108826,],\n",
+      "[-0.43599408864974976,],\n",
+      "[0.954651415348053,],\n",
+      "[0.8184202909469604,],\n",
+      "[-0.9386465549468994,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2524396777153015,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.16811853647232056,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505271792411804,],\n",
+      "[-0.13039612770080566,],\n",
+      "[0.5422419309616089,],\n",
+      "[-0.9990394711494446,],\n",
+      "[-0.04470048472285271,],\n",
+      "[0.45201024413108826,],\n",
+      "[-0.43599408864974976,],\n",
+      "[0.954651415348053,],\n",
+      "[0.8184202909469604,],\n",
+      "[-0.9386465549468994,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00010636015213094652,],\n",
+      "[5.2168528782203794e-05,],\n",
+      "[0.000541498011443764,],\n",
+      "[-0.0001429366529919207,],\n",
+      "[-0.5133551955223083,],\n",
+      "[-0.05214408412575722,],\n",
+      "[-0.9031083583831787,],\n",
+      "[-0.2917688190937042,],\n",
+      "[-0.08261612802743912,],\n",
+      "[0.0028133844025433064,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2524396777153015,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.8200452923774719,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.65052729845047,],\n",
+      "[-0.13039618730545044,],\n",
+      "[0.5422413945198059,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.04418712854385376,],\n",
+      "[0.4520623981952667,],\n",
+      "[-0.4350909888744354,],\n",
+      "[0.9549431800842285,],\n",
+      "[0.8185029029846191,],\n",
+      "[-0.9386493563652039,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25325971841812134,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1403256058692932,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.65052729845047,],\n",
+      "[-0.13039618730545044,],\n",
+      "[0.5422413945198059,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.04418712854385376,],\n",
+      "[0.4520623981952667,],\n",
+      "[-0.4350909888744354,],\n",
+      "[0.9549431800842285,],\n",
+      "[0.8185029029846191,],\n",
+      "[-0.9386493563652039,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.152018952183425e-05,],\n",
+      "[3.5921526432503015e-05,],\n",
+      "[0.0003957444569095969,],\n",
+      "[-0.00010714885138440877,],\n",
+      "[-0.5089207291603088,],\n",
+      "[-0.051612235605716705,],\n",
+      "[-0.8895718455314636,],\n",
+      "[-0.29256772994995117,],\n",
+      "[-0.08261396735906601,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25325971841812134,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7492009997367859,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505273580551147,],\n",
+      "[-0.13039621710777283,],\n",
+      "[0.5422409772872925,],\n",
+      "[-0.9990392327308655,],\n",
+      "[-0.04367820918560028,],\n",
+      "[0.4521140158176422,],\n",
+      "[-0.4342014193534851,],\n",
+      "[0.9552357196807861,],\n",
+      "[0.8185855150222778,],\n",
+      "[-0.9386521577835083,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2540089190006256,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.004706420470029116,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505273580551147,],\n",
+      "[-0.13039621710777283,],\n",
+      "[0.5422409772872925,],\n",
+      "[-0.9990392327308655,],\n",
+      "[-0.04367820918560028,],\n",
+      "[0.4521140158176422,],\n",
+      "[-0.4342014193534851,],\n",
+      "[0.9552357196807861,],\n",
+      "[0.8185855150222778,],\n",
+      "[-0.9386521577835083,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-9.640131793275941e-06,],\n",
+      "[4.9590244088904e-06,],\n",
+      "[5.797594349132851e-05,],\n",
+      "[-1.609995524631813e-05,],\n",
+      "[-0.4848536550998688,],\n",
+      "[-0.04708011820912361,],\n",
+      "[-0.7780134677886963,],\n",
+      "[-0.29326605796813965,],\n",
+      "[-0.08261103928089142,],\n",
+      "[0.0028133802115917206,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2540089190006256,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.13720686733722687,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505273580551147,],\n",
+      "[-0.13039621710777283,],\n",
+      "[0.5422409176826477,],\n",
+      "[-0.9990392327308655,],\n",
+      "[-0.043193355202674866,],\n",
+      "[0.45216110348701477,],\n",
+      "[-0.43342339992523193,],\n",
+      "[0.955528974533081,],\n",
+      "[0.8186681270599365,],\n",
+      "[-0.9386549592018127,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25414612889289856,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10372734814882278,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505273580551147,],\n",
+      "[-0.13039621710777283,],\n",
+      "[0.5422409176826477,],\n",
+      "[-0.9990392327308655,],\n",
+      "[-0.043193355202674866,],\n",
+      "[0.45216110348701477,],\n",
+      "[-0.43342339992523193,],\n",
+      "[0.955528974533081,],\n",
+      "[0.8186681270599365,],\n",
+      "[-0.9386549592018127,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.330820982228033e-05,],\n",
+      "[-1.7581422071089037e-05,],\n",
+      "[-0.00021782072144560516,],\n",
+      "[6.200969801284373e-05,],\n",
+      "[-0.45736658573150635,],\n",
+      "[-0.041481394320726395,],\n",
+      "[-0.6372496485710144,],\n",
+      "[-0.29336151480674744,],\n",
+      "[-0.08260881155729294,],\n",
+      "[0.002813379978761077,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25414612889289856,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6441349983215332,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.65052729845047,],\n",
+      "[-0.13039620220661163,],\n",
+      "[0.5422411561012268,],\n",
+      "[-0.9990392923355103,],\n",
+      "[-0.0427359901368618,],\n",
+      "[0.45220258831977844,],\n",
+      "[-0.43278613686561584,],\n",
+      "[0.9558223485946655,],\n",
+      "[0.8187507390975952,],\n",
+      "[-0.9386577606201172,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2535019814968109,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3136410415172577,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.65052729845047,],\n",
+      "[-0.13039620220661163,],\n",
+      "[0.5422411561012268,],\n",
+      "[-0.9990392923355103,],\n",
+      "[-0.0427359901368618,],\n",
+      "[0.45220258831977844,],\n",
+      "[-0.43278613686561584,],\n",
+      "[0.9558223485946655,],\n",
+      "[0.8187507390975952,],\n",
+      "[-0.9386577606201172,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[4.2626528738765046e-05,],\n",
+      "[-2.314171433681622e-05,],\n",
+      "[-0.0003033075190614909,],\n",
+      "[8.845891716191545e-05,],\n",
+      "[-0.443144828081131,],\n",
+      "[-0.038248032331466675,],\n",
+      "[-0.5514288544654846,],\n",
+      "[-0.29270556569099426,],\n",
+      "[-0.08260804414749146,],\n",
+      "[0.002813380677253008,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2535019814968109,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.1200735569000244,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039617240428925,],\n",
+      "[0.5422414541244507,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.042292844504117966,],\n",
+      "[0.45224082469940186,],\n",
+      "[-0.4322347044944763,],\n",
+      "[0.9561150670051575,],\n",
+      "[0.8188333511352539,],\n",
+      "[-0.9386605620384216,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25238192081451416,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2505861222743988,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039617240428925,],\n",
+      "[0.5422414541244507,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.042292844504117966,],\n",
+      "[0.45224082469940186,],\n",
+      "[-0.4322347044944763,],\n",
+      "[0.9561150670051575,],\n",
+      "[0.8188333511352539,],\n",
+      "[-0.9386605620384216,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.804117502819281e-05,],\n",
+      "[-1.5679876014473848e-05,],\n",
+      "[-0.00021718113566748798,],\n",
+      "[6.48634450044483e-05,],\n",
+      "[-0.4490736126899719,],\n",
+      "[-0.03913938254117966,],\n",
+      "[-0.5717135667800903,],\n",
+      "[-0.2916005551815033,],\n",
+      "[-0.08260875940322876,],\n",
+      "[0.002813381375744939,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25238192081451416,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0011729001998901,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04184377193450928,],\n",
+      "[0.4522799551486969,],\n",
+      "[-0.43166297674179077,],\n",
+      "[0.9564066529273987,],\n",
+      "[0.8189159631729126,],\n",
+      "[-0.9386633634567261,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2513807415962219,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03229731693863869,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04184377193450928,],\n",
+      "[0.4522799551486969,],\n",
+      "[-0.43166297674179077,],\n",
+      "[0.9564066529273987,],\n",
+      "[0.8189159631729126,],\n",
+      "[-0.9386633634567261,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.408882993331645e-06,],\n",
+      "[-4.265504685463384e-06,],\n",
+      "[-6.24531094217673e-05,],\n",
+      "[1.9102648366242647e-05,],\n",
+      "[-0.4693303406238556,],\n",
+      "[-0.04330570995807648,],\n",
+      "[-0.6866486668586731,],\n",
+      "[-0.290620356798172,],\n",
+      "[-0.08260948956012726,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2513807415962219,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3594287633895874,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422417521476746,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04137444123625755,],\n",
+      "[0.4523232579231262,],\n",
+      "[-0.43097633123397827,],\n",
+      "[0.9566972851753235,],\n",
+      "[0.8189985752105713,],\n",
+      "[-0.9386661648750305,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2510213255882263,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04306073486804962,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422417521476746,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04137444123625755,],\n",
+      "[0.4523232579231262,],\n",
+      "[-0.43097633123397827,],\n",
+      "[0.9566972851753235,],\n",
+      "[0.8189985752105713,],\n",
+      "[-0.9386661648750305,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.2959779825177975e-06,],\n",
+      "[3.7248989883664763e-06,],\n",
+      "[5.7729921536520123e-05,],\n",
+      "[-1.809403875085991e-05,],\n",
+      "[-0.4906906485557556,],\n",
+      "[-0.04806092008948326,],\n",
+      "[-0.8261707425117493,],\n",
+      "[-0.2902788519859314,],\n",
+      "[-0.08261022716760635,],\n",
+      "[0.00281338207423687,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2510213255882263,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.41502124071121216,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04088374972343445,],\n",
+      "[0.4523713290691376,],\n",
+      "[-0.4301501512527466,],\n",
+      "[0.9569875597953796,],\n",
+      "[0.81908118724823,],\n",
+      "[-0.938668966293335,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25143635272979736,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1810634732246399,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.9990394115447998,],\n",
+      "[-0.04088374972343445,],\n",
+      "[0.4523713290691376,],\n",
+      "[-0.4301501512527466,],\n",
+      "[0.9569875597953796,],\n",
+      "[0.81908118724823,],\n",
+      "[-0.938668966293335,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-9.501482054474764e-06,],\n",
+      "[5.763243734691059e-06,],\n",
+      "[9.470582153880969e-05,],\n",
+      "[-3.0436285669566132e-05,],\n",
+      "[-0.5010535717010498,],\n",
+      "[-0.05060333386063576,],\n",
+      "[-0.9037569761276245,],\n",
+      "[-0.2906968295574188,],\n",
+      "[-0.08261097967624664,],\n",
+      "[0.00281338207423687,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25143635272979736,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.8510304689407349,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415733337402,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.04038269445300102,],\n",
+      "[0.4524219334125519,],\n",
+      "[-0.4292463958263397,],\n",
+      "[0.9572782516479492,],\n",
+      "[0.8191637992858887,],\n",
+      "[-0.9386717677116394,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2522873878479004,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11665084213018417,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415733337402,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.04038269445300102,],\n",
+      "[0.4524219334125519,],\n",
+      "[-0.4292463958263397,],\n",
+      "[0.9572782516479492,],\n",
+      "[0.8191637992858887,],\n",
+      "[-0.9386717677116394,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.612689164991025e-06,],\n",
+      "[3.48582193510083e-06,],\n",
+      "[6.079099330236204e-05,],\n",
+      "[-2.0040053641423583e-05,],\n",
+      "[-0.4958886504173279,],\n",
+      "[-0.04960493743419647,],\n",
+      "[-0.8725844025611877,],\n",
+      "[-0.2915448248386383,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2522873878479004,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6830832958221436,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415137290955,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03988680616021156,],\n",
+      "[0.452471524477005,],\n",
+      "[-0.4283738136291504,],\n",
+      "[0.9575697779655457,],\n",
+      "[0.8192464113235474,],\n",
+      "[-0.9386745691299438,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25297045707702637,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([3.9712231227895245e-05,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415137290955,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03988680616021156,],\n",
+      "[0.452471524477005,],\n",
+      "[-0.4283738136291504,],\n",
+      "[0.9575697779655457,],\n",
+      "[0.8192464113235474,],\n",
+      "[-0.9386745691299438,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.621500230925449e-08,],\n",
+      "[4.848956081104916e-08,],\n",
+      "[8.9711841155804e-07,],\n",
+      "[-3.033118503026344e-07,],\n",
+      "[-0.4799652397632599,],\n",
+      "[-0.04589144140481949,],\n",
+      "[-0.7547435164451599,],\n",
+      "[-0.292218953371048,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25297045707702637,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.012603516690433025,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415137290955,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03940683975815773,],\n",
+      "[0.45251742005348206,],\n",
+      "[-0.427619069814682,],\n",
+      "[0.9578620195388794,],\n",
+      "[0.819329023361206,],\n",
+      "[-0.9386773705482483,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25298306345939636,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1420767605304718,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415137290955,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03940683975815773,],\n",
+      "[0.45251742005348206,],\n",
+      "[-0.427619069814682,],\n",
+      "[0.9578620195388794,],\n",
+      "[0.819329023361206,],\n",
+      "[-0.9386773705482483,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.3549879390193382e-06,],\n",
+      "[-2.1910175291850464e-06,],\n",
+      "[-4.294257814763114e-05,],\n",
+      "[1.4881794413668104e-05,],\n",
+      "[-0.46384578943252563,],\n",
+      "[-0.04185907542705536,],\n",
+      "[-0.6224113702774048,],\n",
+      "[-0.29222559928894043,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25298306345939636,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7538617253303528,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415733337402,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03894299268722534,],\n",
+      "[0.45255929231643677,],\n",
+      "[-0.426996648311615,],\n",
+      "[0.9581542611122131,],\n",
+      "[0.8194116353988647,],\n",
+      "[-0.9386801719665527,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2522292137145996,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3312968313694,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422415733337402,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03894299268722534,],\n",
+      "[0.45255929231643677,],\n",
+      "[-0.426996648311615,],\n",
+      "[0.9581542611122131,],\n",
+      "[0.8194116353988647,],\n",
+      "[-0.9386801719665527,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.7703839552705176e-06,],\n",
+      "[-2.5333085886813933e-06,],\n",
+      "[-5.251195398159325e-05,],\n",
+      "[1.8640761481947266e-05,],\n",
+      "[-0.4570728838443756,],\n",
+      "[-0.03993135318160057,],\n",
+      "[-0.5527490973472595,],\n",
+      "[-0.2914717495441437,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2522292137145996,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.1511682271957397,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.542241632938385,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03848591819405556,],\n",
+      "[0.4525992274284363,],\n",
+      "[-0.4264439046382904,],\n",
+      "[0.9584457278251648,],\n",
+      "[0.8194942474365234,],\n",
+      "[-0.9386829733848572,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2510780394077301,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2198541760444641,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.542241632938385,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03848591819405556,],\n",
+      "[0.4525992274284363,],\n",
+      "[-0.4264439046382904,],\n",
+      "[0.9584457278251648,],\n",
+      "[0.8194942474365234,],\n",
+      "[-0.9386829733848572,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.2604426703765057e-06,],\n",
+      "[-1.5644338873244124e-06,],\n",
+      "[-3.426795228733681e-05,],\n",
+      "[1.2456049262254965e-05,],\n",
+      "[-0.46247753500938416,],\n",
+      "[-0.041074238717556,],\n",
+      "[-0.587500274181366,],\n",
+      "[-0.29032060503959656,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2510780394077301,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.937771737575531,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.038023442029953,],\n",
+      "[0.4526402950286865,],\n",
+      "[-0.42585641145706177,],\n",
+      "[0.9587360620498657,],\n",
+      "[0.8195768594741821,],\n",
+      "[-0.9386857748031616,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2501402795314789,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.014800670556724072,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.038023442029953,],\n",
+      "[0.4526402950286865,],\n",
+      "[-0.42585641145706177,],\n",
+      "[0.9587360620498657,],\n",
+      "[0.8195768594741821,],\n",
+      "[-0.9386857748031616,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[4.3163421992176154e-07,],\n",
+      "[-3.075112431361049e-07,],\n",
+      "[-7.121204816940008e-06,],\n",
+      "[2.6510031148063717e-06,],\n",
+      "[-0.4754222631454468,],\n",
+      "[-0.04435121640563011,],\n",
+      "[-0.7079465985298157,],\n",
+      "[-0.2893888056278229,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2501402795314789,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.24331627786159515,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03754802048206329,],\n",
+      "[0.4526846408843994,],\n",
+      "[-0.42514845728874207,],\n",
+      "[0.9590254426002502,],\n",
+      "[0.8196594715118408,],\n",
+      "[-0.9386885762214661,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.24989695847034454,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.06487461924552917,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03754802048206329,],\n",
+      "[0.4526846408843994,],\n",
+      "[-0.42514845728874207,],\n",
+      "[0.9590254426002502,],\n",
+      "[0.8196594715118408,],\n",
+      "[-0.9386885762214661,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.650603836533264e-07,],\n",
+      "[4.867253551310569e-07,],\n",
+      "[1.1933774658245966e-05,],\n",
+      "[-4.552639438770711e-06,],\n",
+      "[-0.48765310645103455,],\n",
+      "[-0.04767552763223648,],\n",
+      "[-0.8400094509124756,],\n",
+      "[-0.2891484797000885,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.24989695847034454,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5094103813171387,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.037060368806123734,],\n",
+      "[0.4527323246002197,],\n",
+      "[-0.4243084490299225,],\n",
+      "[0.9593145847320557,],\n",
+      "[0.8197420835494995,],\n",
+      "[-0.9386913776397705,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2504063546657562,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1857844889163971,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.037060368806123734,],\n",
+      "[0.4527323246002197,],\n",
+      "[-0.4243084490299225,],\n",
+      "[0.9593145847320557,],\n",
+      "[0.8197420835494995,],\n",
+      "[-0.9386913776397705,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.282784165203338e-07,],\n",
+      "[6.213047072378686e-07,],\n",
+      "[1.615433211554773e-05,],\n",
+      "[-6.319565727608278e-06,],\n",
+      "[-0.49243003129959106,],\n",
+      "[-0.04912285506725311,],\n",
+      "[-0.9001957774162292,],\n",
+      "[-0.2896578907966614,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2504063546657562,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.862054705619812,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03656793758273125,],\n",
+      "[0.45278143882751465,],\n",
+      "[-0.42340824007987976,],\n",
+      "[0.9596042633056641,],\n",
+      "[0.8198246955871582,],\n",
+      "[-0.938694179058075,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2512684166431427,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08998353779315948,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03656793758273125,],\n",
+      "[0.45278143882751465,],\n",
+      "[-0.42340824007987976,],\n",
+      "[0.9596042633056641,],\n",
+      "[0.8198246955871582,],\n",
+      "[-0.938694179058075,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-4.242303646151413e-07,],\n",
+      "[3.2581789355390356e-07,],\n",
+      "[8.990295100375079e-06,],\n",
+      "[-3.607602138799848e-06,],\n",
+      "[-0.48800528049468994,],\n",
+      "[-0.048006270080804825,],\n",
+      "[-0.8531227111816406,],\n",
+      "[-0.2905169427394867,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2512684166431427,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5999447107315063,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03607993200421333,],\n",
+      "[0.4528294503688812,],\n",
+      "[-0.4225551187992096,],\n",
+      "[0.9598947763442993,],\n",
+      "[0.8199073076248169,],\n",
+      "[-0.9386969804763794,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2518683671951294,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.003489461727440357,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03607993200421333,],\n",
+      "[0.4528294503688812,],\n",
+      "[-0.4225551187992096,],\n",
+      "[0.9598947763442993,],\n",
+      "[0.8199073076248169,],\n",
+      "[-0.9386969804763794,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[6.148190578869617e-08,],\n",
+      "[-4.838430456288734e-08,],\n",
+      "[-1.416049030922295e-06,],\n",
+      "[5.827338327435427e-07,],\n",
+      "[-0.4780634343624115,],\n",
+      "[-0.04518555477261543,],\n",
+      "[-0.7322211265563965,],\n",
+      "[-0.2911168932914734,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2518683671951294,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.11814315617084503,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.035601869225502014,],\n",
+      "[0.45287463068962097,],\n",
+      "[-0.4218229055404663,],\n",
+      "[0.9601858854293823,],\n",
+      "[0.8199899196624756,],\n",
+      "[-0.9386997818946838,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.25175023078918457,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.18360266089439392,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.035601869225502014,],\n",
+      "[0.45287463068962097,],\n",
+      "[-0.4218229055404663,],\n",
+      "[0.9601858854293823,],\n",
+      "[0.8199899196624756,],\n",
+      "[-0.9386997818946838,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.2821387208059605e-07,],\n",
+      "[-2.6523520091359387e-07,],\n",
+      "[-8.220880772569217e-06,],\n",
+      "[3.4672391393542057e-06,],\n",
+      "[-0.46908363699913025,],\n",
+      "[-0.0424647331237793,],\n",
+      "[-0.6101729273796082,],\n",
+      "[-0.29099875688552856,],\n",
+      "[-0.08261024206876755,],\n",
+      "[0.002813381841406226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.25175023078918457,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8569777607917786,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03513278439640999,],\n",
+      "[0.45291709899902344,],\n",
+      "[-0.42121273279190063,],\n",
+      "[0.9604768753051758,],\n",
+      "[0.8200725317001343,],\n",
+      "[-0.9387025833129883,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2508932650089264,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [4/10], Loss: 0.1836\n",
+      "loss_antes tensor([0.7371393442153931,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6505272388458252,],\n",
+      "[-0.13039615750312805,],\n",
+      "[0.5422416925430298,],\n",
+      "[-0.999039351940155,],\n",
+      "[-0.03513278439640999,],\n",
+      "[0.45291709899902344,],\n",
+      "[-0.42121273279190063,],\n",
+      "[0.9604768753051758,],\n",
+      "[0.8200725317001343,],\n",
+      "[-0.9387025833129883,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3514309525489807,],\n",
+      "[-0.1430390626192093,],\n",
+      "[-0.4543296694755554,],\n",
+      "[0.07448316365480423,],\n",
+      "[-0.2737724184989929,],\n",
+      "[-0.02696884423494339,],\n",
+      "[-0.44261306524276733,],\n",
+      "[-0.19041219353675842,],\n",
+      "[-0.05249946564435959,],\n",
+      "[0.001837005140259862,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2508932650089264,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7171374559402466,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6501758098602295,],\n",
+      "[-0.1302531212568283,],\n",
+      "[0.5426959991455078,],\n",
+      "[-0.9991138577461243,],\n",
+      "[-0.03485901281237602,],\n",
+      "[0.4529440701007843,],\n",
+      "[-0.4207701086997986,],\n",
+      "[0.9606673121452332,],\n",
+      "[0.8201250433921814,],\n",
+      "[-0.9387044310569763,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.24917612969875336,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.723792314529419,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6501758098602295,],\n",
+      "[-0.1302531212568283,],\n",
+      "[0.5426959991455078,],\n",
+      "[-0.9991138577461243,],\n",
+      "[-0.03485901281237602,],\n",
+      "[0.4529440701007843,],\n",
+      "[-0.4207701086997986,],\n",
+      "[0.9606673121452332,],\n",
+      "[0.8201250433921814,],\n",
+      "[-0.9387044310569763,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.37474557757377625,],\n",
+      "[-0.14188139140605927,],\n",
+      "[-0.41997817158699036,],\n",
+      "[0.06670809537172318,],\n",
+      "[-0.2756737172603607,],\n",
+      "[-0.026301689445972443,],\n",
+      "[-0.4436373710632324,],\n",
+      "[-0.20089367032051086,],\n",
+      "[-0.048812299966812134,],\n",
+      "[0.002360919024795294,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.24917612969875336,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.7015209197998047,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6498010754585266,],\n",
+      "[-0.13011124730110168,],\n",
+      "[0.5431159734725952,],\n",
+      "[-0.9991805553436279,],\n",
+      "[-0.03458333760499954,],\n",
+      "[0.45297038555145264,],\n",
+      "[-0.42032647132873535,],\n",
+      "[0.9608681797981262,],\n",
+      "[0.8201738595962524,],\n",
+      "[-0.9387068152427673,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.24747461080551147,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1125712394714355,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6498010754585266,],\n",
+      "[-0.13011124730110168,],\n",
+      "[0.5431159734725952,],\n",
+      "[-0.9991805553436279,],\n",
+      "[-0.03458333760499954,],\n",
+      "[0.45297038555145264,],\n",
+      "[-0.42032647132873535,],\n",
+      "[0.9608681797981262,],\n",
+      "[0.8201738595962524,],\n",
+      "[-0.9387068152427673,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.47777873277664185,],\n",
+      "[-0.1694008857011795,],\n",
+      "[-0.47538596391677856,],\n",
+      "[0.07373511046171188,],\n",
+      "[-0.22786098718643188,],\n",
+      "[-0.0212094709277153,],\n",
+      "[-0.3680896461009979,],\n",
+      "[-0.19223439693450928,],\n",
+      "[-0.038274869322776794,],\n",
+      "[0.0026440832298249006,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.24747461080551147,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1095666885375977,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6493232846260071,],\n",
+      "[-0.12994185090065002,],\n",
+      "[0.5435913801193237,],\n",
+      "[-0.9992542862892151,],\n",
+      "[-0.03435547649860382,],\n",
+      "[0.4529916048049927,],\n",
+      "[-0.4199583828449249,],\n",
+      "[0.9610604047775269,],\n",
+      "[0.8202121257781982,],\n",
+      "[-0.9387094378471375,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.24536503851413727,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.5711164474487305,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6493232846260071,],\n",
+      "[-0.12994185090065002,],\n",
+      "[0.5435913801193237,],\n",
+      "[-0.9992542862892151,],\n",
+      "[-0.03435547649860382,],\n",
+      "[0.4529916048049927,],\n",
+      "[-0.4199583828449249,],\n",
+      "[0.9610604047775269,],\n",
+      "[0.8202121257781982,],\n",
+      "[-0.9387094378471375,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.5569821000099182,],\n",
+      "[-0.1869160681962967,],\n",
+      "[-0.505743145942688,],\n",
+      "[0.07713517546653748,],\n",
+      "[-0.18264959752559662,],\n",
+      "[-0.016307346522808075,],\n",
+      "[-0.2940630316734314,],\n",
+      "[-0.18858784437179565,],\n",
+      "[-0.029883116483688354,],\n",
+      "[0.002990629058331251,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.24536503851413727,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.5068836212158203,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.648766279220581,],\n",
+      "[-0.1297549307346344,],\n",
+      "[0.5440971255302429,],\n",
+      "[-0.9993314146995544,],\n",
+      "[-0.03417282551527023,],\n",
+      "[0.45300790667533875,],\n",
+      "[-0.41966432332992554,],\n",
+      "[0.9612489938735962,],\n",
+      "[0.8202419877052307,],\n",
+      "[-0.9387124180793762,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2428581565618515,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.5283877849578857,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.648766279220581,],\n",
+      "[-0.1297549307346344,],\n",
+      "[0.5440971255302429,],\n",
+      "[-0.9993314146995544,],\n",
+      "[-0.03417282551527023,],\n",
+      "[0.45300790667533875,],\n",
+      "[-0.41966432332992554,],\n",
+      "[0.9612489938735962,],\n",
+      "[0.8202419877052307,],\n",
+      "[-0.9387124180793762,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.514426589012146,],\n",
+      "[-0.1655067354440689,],\n",
+      "[-0.4386959373950958,],\n",
+      "[0.06623160094022751,],\n",
+      "[-0.1891956627368927,],\n",
+      "[-0.015920769423246384,],\n",
+      "[-0.29826632142066956,],\n",
+      "[-0.20264087617397308,],\n",
+      "[-0.03254454582929611,],\n",
+      "[0.003492067102342844,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2428581565618515,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.472560167312622,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6482518315315247,],\n",
+      "[-0.12958942353725433,],\n",
+      "[0.5445358157157898,],\n",
+      "[-0.9993976354598999,],\n",
+      "[-0.03398362919688225,],\n",
+      "[0.4530238211154938,],\n",
+      "[-0.4193660616874695,],\n",
+      "[0.9614516496658325,],\n",
+      "[0.8202745318412781,],\n",
+      "[-0.938715934753418,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.24038559198379517,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8637006878852844,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6482518315315247,],\n",
+      "[-0.12958942353725433,],\n",
+      "[0.5445358157157898,],\n",
+      "[-0.9993976354598999,],\n",
+      "[-0.03398362919688225,],\n",
+      "[0.4530238211154938,],\n",
+      "[-0.4193660616874695,],\n",
+      "[0.9614516496658325,],\n",
+      "[0.8202745318412781,],\n",
+      "[-0.938715934753418,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.34661489725112915,],\n",
+      "[-0.10829398781061172,],\n",
+      "[-0.28542885184288025,],\n",
+      "[0.0429387129843235,],\n",
+      "[-0.26254603266716003,],\n",
+      "[-0.021255042403936386,],\n",
+      "[-0.40823763608932495,],\n",
+      "[-0.2306896150112152,],\n",
+      "[-0.04659931734204292,],\n",
+      "[0.004023045301437378,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.24038559198379517,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8587106466293335,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6479052305221558,],\n",
+      "[-0.12948113679885864,],\n",
+      "[0.5448212623596191,],\n",
+      "[-0.9994405508041382,],\n",
+      "[-0.03372108191251755,],\n",
+      "[0.4530450701713562,],\n",
+      "[-0.41895782947540283,],\n",
+      "[0.9616823196411133,],\n",
+      "[0.8203211426734924,],\n",
+      "[-0.9387199282646179,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23852688074111938,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.19525720179080963,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6479052305221558,],\n",
+      "[-0.12948113679885864,],\n",
+      "[0.5448212623596191,],\n",
+      "[-0.9994405508041382,],\n",
+      "[-0.03372108191251755,],\n",
+      "[0.4530450701713562,],\n",
+      "[-0.41895782947540283,],\n",
+      "[0.9616823196411133,],\n",
+      "[0.8203211426734924,],\n",
+      "[-0.9387199282646179,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.14204828441143036,],\n",
+      "[-0.04357573017477989,],\n",
+      "[-0.11582156270742416,],\n",
+      "[0.01747106947004795,],\n",
+      "[-0.37485557794570923,],\n",
+      "[-0.029972977936267853,],\n",
+      "[-0.5847826600074768,],\n",
+      "[-0.2599247395992279,],\n",
+      "[-0.06412062048912048,],\n",
+      "[0.004413783084601164,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23852688074111938,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8837581276893616,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.647763192653656,],\n",
+      "[-0.12943756580352783,],\n",
+      "[0.5449370741844177,],\n",
+      "[-0.9994580149650574,],\n",
+      "[-0.03334622457623482,],\n",
+      "[0.4530750513076782,],\n",
+      "[-0.4183730483055115,],\n",
+      "[0.9619422554969788,],\n",
+      "[0.8203852772712708,],\n",
+      "[-0.9387243390083313,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23764312267303467,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.267035145247064e-06,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.647763192653656,],\n",
+      "[-0.12943756580352783,],\n",
+      "[0.5449370741844177,],\n",
+      "[-0.9994580149650574,],\n",
+      "[-0.03334622457623482,],\n",
+      "[0.4530750513076782,],\n",
+      "[-0.4183730483055115,],\n",
+      "[0.9619422554969788,],\n",
+      "[0.8203852772712708,],\n",
+      "[-0.9387243390083313,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0003016701666638255,],\n",
+      "[-9.164163930108771e-05,],\n",
+      "[-0.00024882328580133617,],\n",
+      "[3.7839934520889074e-05,],\n",
+      "[-0.4735064208507538,],\n",
+      "[-0.03794558346271515,],\n",
+      "[-0.7453377842903137,],\n",
+      "[-0.2791528105735779,],\n",
+      "[-0.07631070166826248,],\n",
+      "[0.0045994604006409645,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23764312267303467,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.0022512529976665974,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6477628946304321,],\n",
+      "[-0.12943747639656067,],\n",
+      "[0.5449373126029968,],\n",
+      "[-0.9994580745697021,],\n",
+      "[-0.03287271782755852,],\n",
+      "[0.45311298966407776,],\n",
+      "[-0.4176277220249176,],\n",
+      "[0.9622213840484619,],\n",
+      "[0.8204615712165833,],\n",
+      "[-0.938728928565979,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2376408725976944,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.034220948815345764,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6477628946304321,],\n",
+      "[-0.12943747639656067,],\n",
+      "[0.5449373126029968,],\n",
+      "[-0.9994580745697021,],\n",
+      "[-0.03287271782755852,],\n",
+      "[0.45311298966407776,],\n",
+      "[-0.4176277220249176,],\n",
+      "[0.9622213840484619,],\n",
+      "[0.8204615712165833,],\n",
+      "[-0.938728928565979,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.04021427780389786,],\n",
+      "[0.012179398909211159,],\n",
+      "[0.03415193036198616,],\n",
+      "[-0.005258030723780394,],\n",
+      "[-0.5137218832969666,],\n",
+      "[-0.04147711396217346,],\n",
+      "[-0.813643753528595,],\n",
+      "[-0.28487277030944824,],\n",
+      "[-0.07979565858840942,],\n",
+      "[0.004631139803677797,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2376408725976944,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.36997801065444946,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6478031277656555,],\n",
+      "[-0.12944965064525604,],\n",
+      "[0.5449031591415405,],\n",
+      "[-0.9994528293609619,],\n",
+      "[-0.03235899657011032,],\n",
+      "[0.45315447449684143,],\n",
+      "[-0.41681408882141113,],\n",
+      "[0.9625062346458435,],\n",
+      "[0.8205413818359375,],\n",
+      "[-0.9387335777282715,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23801085352897644,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.00293249124661088,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6478031277656555,],\n",
+      "[-0.12944965064525604,],\n",
+      "[0.5449031591415405,],\n",
+      "[-0.9994528293609619,],\n",
+      "[-0.03235899657011032,],\n",
+      "[0.45315447449684143,],\n",
+      "[-0.41681408882141113,],\n",
+      "[0.9625062346458435,],\n",
+      "[0.8205413818359375,],\n",
+      "[-0.9387335777282715,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.009337124414741993,],\n",
+      "[0.0028374437242746353,],\n",
+      "[0.008282900787889957,],\n",
+      "[-0.0012949843658134341,],\n",
+      "[-0.48532718420028687,],\n",
+      "[-0.03918362408876419,],\n",
+      "[-0.7658417224884033,],\n",
+      "[-0.2808997929096222,],\n",
+      "[-0.07718196511268616,],\n",
+      "[0.004599194973707199,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23801085352897644,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.10830515623092651,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6478124856948853,],\n",
+      "[-0.12945248186588287,],\n",
+      "[0.5448948740959167,],\n",
+      "[-0.9994515180587769,],\n",
+      "[-0.031873669475317,],\n",
+      "[0.45319366455078125,],\n",
+      "[-0.4160482585430145,],\n",
+      "[0.9627871513366699,],\n",
+      "[0.8206185698509216,],\n",
+      "[-0.9387381672859192,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23811915516853333,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07396981120109558,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6478124856948853,],\n",
+      "[-0.12945248186588287,],\n",
+      "[0.5448948740959167,],\n",
+      "[-0.9994515180587769,],\n",
+      "[-0.031873669475317,],\n",
+      "[0.45319366455078125,],\n",
+      "[-0.4160482585430145,],\n",
+      "[0.9627871513366699,],\n",
+      "[0.8206185698509216,],\n",
+      "[-0.9387381672859192,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.03651640936732292,],\n",
+      "[-0.011211820878088474,],\n",
+      "[-0.03423565253615379,],\n",
+      "[0.0054476941004395485,],\n",
+      "[-0.4178314805030823,],\n",
+      "[-0.033022690564394,],\n",
+      "[-0.6456075310707092,],\n",
+      "[-0.2741037607192993,],\n",
+      "[-0.07334686070680618,],\n",
+      "[0.004569435026496649,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23811915516853333,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5439467430114746,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6477759480476379,],\n",
+      "[-0.1294412761926651,],\n",
+      "[0.5449290871620178,],\n",
+      "[-0.9994569420814514,],\n",
+      "[-0.031455837190151215,],\n",
+      "[0.4532266855239868,],\n",
+      "[-0.4154026508331299,],\n",
+      "[0.9630612730979919,],\n",
+      "[0.8206919431686401,],\n",
+      "[-0.9387427568435669,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2375752031803131,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3143860101699829,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6477759480476379,],\n",
+      "[-0.1294412761926651,],\n",
+      "[0.5449290871620178,],\n",
+      "[-0.9994569420814514,],\n",
+      "[-0.031455837190151215,],\n",
+      "[0.4532266855239868,],\n",
+      "[-0.4154026508331299,],\n",
+      "[0.9630612730979919,],\n",
+      "[0.8206919431686401,],\n",
+      "[-0.9387427568435669,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.05775394290685654,],\n",
+      "[-0.018041005358099937,],\n",
+      "[-0.05775866284966469,],\n",
+      "[0.00937072653323412,],\n",
+      "[-0.3612804114818573,],\n",
+      "[-0.027375422418117523,],\n",
+      "[-0.5384925603866577,],\n",
+      "[-0.26989683508872986,],\n",
+      "[-0.07160915434360504,],\n",
+      "[0.004564919974654913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2375752031803131,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.1214019060134888,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6477181911468506,],\n",
+      "[-0.12942323088645935,],\n",
+      "[0.5449868440628052,],\n",
+      "[-0.9994663000106812,],\n",
+      "[-0.03109455667436123,],\n",
+      "[0.4532540738582611,],\n",
+      "[-0.4148641526699066,],\n",
+      "[0.9633311629295349,],\n",
+      "[0.8207635283470154,],\n",
+      "[-0.9387473464012146,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2364538013935089,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3702930510044098,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6477181911468506,],\n",
+      "[-0.12942323088645935,],\n",
+      "[0.5449868440628052,],\n",
+      "[-0.9994663000106812,],\n",
+      "[-0.03109455667436123,],\n",
+      "[0.4532540738582611,],\n",
+      "[-0.4148641526699066,],\n",
+      "[0.9633311629295349,],\n",
+      "[0.8207635283470154,],\n",
+      "[-0.9387473464012146,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04753384366631508,],\n",
+      "[-0.015190435573458672,],\n",
+      "[-0.05105670168995857,],\n",
+      "[0.008458138443529606,],\n",
+      "[-0.355071097612381,],\n",
+      "[-0.02617274597287178,],\n",
+      "[-0.5201268792152405,],\n",
+      "[-0.2699896991252899,],\n",
+      "[-0.07245129346847534,],\n",
+      "[0.004579703789204359,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2364538013935089,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.217035174369812,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476706862449646,],\n",
+      "[-0.12940804660320282,],\n",
+      "[0.5450379252433777,],\n",
+      "[-0.9994747638702393,],\n",
+      "[-0.03073948621749878,],\n",
+      "[0.4532802402973175,],\n",
+      "[-0.41434401273727417,],\n",
+      "[0.9636011719703674,],\n",
+      "[0.8208360075950623,],\n",
+      "[-0.9387519359588623,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2352367639541626,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13221073150634766,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476706862449646,],\n",
+      "[-0.12940804660320282,],\n",
+      "[0.5450379252433777,],\n",
+      "[-0.9994747638702393,],\n",
+      "[-0.03073948621749878,],\n",
+      "[0.4532802402973175,],\n",
+      "[-0.41434401273727417,],\n",
+      "[0.9636011719703674,],\n",
+      "[0.8208360075950623,],\n",
+      "[-0.9387519359588623,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.02136050909757614,],\n",
+      "[-0.007003498263657093,],\n",
+      "[-0.024755533784627914,],\n",
+      "[0.00419275090098381,],\n",
+      "[-0.4046832323074341,],\n",
+      "[-0.030302515253424644,],\n",
+      "[-0.6100919246673584,],\n",
+      "[-0.2729955017566681,],\n",
+      "[-0.07454371452331543,],\n",
+      "[0.004598654340952635,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2352367639541626,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7272164821624756,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.647649347782135,],\n",
+      "[-0.1294010430574417,],\n",
+      "[0.5450626611709595,],\n",
+      "[-0.9994789361953735,],\n",
+      "[-0.030334802344441414,],\n",
+      "[0.4533105492591858,],\n",
+      "[-0.41373392939567566,],\n",
+      "[0.9638741612434387,],\n",
+      "[0.8209105730056763,],\n",
+      "[-0.93875652551651,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23450954258441925,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0019009412499144673,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.647649347782135,],\n",
+      "[-0.1294010430574417,],\n",
+      "[0.5450626611709595,],\n",
+      "[-0.9994789361953735,],\n",
+      "[-0.030334802344441414,],\n",
+      "[0.4533105492591858,],\n",
+      "[-0.41373392939567566,],\n",
+      "[0.9638741612434387,],\n",
+      "[0.8209105730056763,],\n",
+      "[-0.93875652551651,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0019151949090883136,],\n",
+      "[0.0006443181191571057,],\n",
+      "[0.00240157637745142,],\n",
+      "[-0.00041627418249845505,],\n",
+      "[-0.4807692766189575,],\n",
+      "[-0.03734869882464409,],\n",
+      "[-0.7604610323905945,],\n",
+      "[-0.27650249004364014,],\n",
+      "[-0.07636477798223495,],\n",
+      "[0.004611185751855373,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23450954258441925,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.08719956129789352,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476512551307678,],\n",
+      "[-0.12940168380737305,],\n",
+      "[0.5450602769851685,],\n",
+      "[-0.9994785189628601,],\n",
+      "[-0.029854033142328262,],\n",
+      "[0.4533478915691376,],\n",
+      "[-0.41297346353530884,],\n",
+      "[0.9641506671905518,],\n",
+      "[0.8209869265556335,],\n",
+      "[-0.9387611150741577,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23459674417972565,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1447117030620575,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476512551307678,],\n",
+      "[-0.12940168380737305,],\n",
+      "[0.5450602769851685,],\n",
+      "[-0.9994785189628601,],\n",
+      "[-0.029854033142328262,],\n",
+      "[0.4533478915691376,],\n",
+      "[-0.41297346353530884,],\n",
+      "[0.9641506671905518,],\n",
+      "[0.8209869265556335,],\n",
+      "[-0.9387611150741577,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.012444978579878807,],\n",
+      "[0.004289716947823763,],\n",
+      "[0.01691252551972866,],\n",
+      "[-0.0030027718748897314,],\n",
+      "[-0.5395131707191467,],\n",
+      "[-0.04332296922802925,],\n",
+      "[-0.8854389190673828,],\n",
+      "[-0.27880561351776123,],\n",
+      "[-0.07716920971870422,],\n",
+      "[0.004614971112459898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23459674417972565,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7608194351196289,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476637125015259,],\n",
+      "[-0.12940597534179688,],\n",
+      "[0.5450433492660522,],\n",
+      "[-0.9994755387306213,],\n",
+      "[-0.029314519837498665,],\n",
+      "[0.4533912241458893,],\n",
+      "[-0.4120880365371704,],\n",
+      "[0.964429497718811,],\n",
+      "[0.8210641145706177,],\n",
+      "[-0.9387657046318054,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23535756766796112,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20596763491630554,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476637125015259,],\n",
+      "[-0.12940597534179688,],\n",
+      "[0.5450433492660522,],\n",
+      "[-0.9994755387306213,],\n",
+      "[-0.029314519837498665,],\n",
+      "[0.4533912241458893,],\n",
+      "[-0.4120880365371704,],\n",
+      "[0.964429497718811,],\n",
+      "[0.8210641145706177,],\n",
+      "[-0.9387657046318054,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.011023963801562786,],\n",
+      "[0.003888155333697796,],\n",
+      "[0.016253717243671417,],\n",
+      "[-0.0029579733964055777,],\n",
+      "[-0.5497504472732544,],\n",
+      "[-0.044904738664627075,],\n",
+      "[-0.9133400917053223,],\n",
+      "[-0.27944669127464294,],\n",
+      "[-0.07705672830343246,],\n",
+      "[0.004613097757101059,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23535756766796112,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9076727032661438,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476747393608093,],\n",
+      "[-0.12940986454486847,],\n",
+      "[0.5450270771980286,],\n",
+      "[-0.9994725584983826,],\n",
+      "[-0.028764769434928894,],\n",
+      "[0.45343613624572754,],\n",
+      "[-0.4111746847629547,],\n",
+      "[0.9647089242935181,],\n",
+      "[0.8211411833763123,],\n",
+      "[-0.9387702941894531,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23626524209976196,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.05217441916465759,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476747393608093,],\n",
+      "[-0.12940986454486847,],\n",
+      "[0.5450270771980286,],\n",
+      "[-0.9994725584983826,],\n",
+      "[-0.028764769434928894,],\n",
+      "[0.45343613624572754,],\n",
+      "[-0.4111746847629547,],\n",
+      "[0.9647089242935181,],\n",
+      "[0.8211411833763123,],\n",
+      "[-0.9387702941894531,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.004109229426831007,],\n",
+      "[0.001483213622123003,],\n",
+      "[0.006579410284757614,],\n",
+      "[-0.0012278452049940825,],\n",
+      "[-0.5104618072509766,],\n",
+      "[-0.04128095135092735,],\n",
+      "[-0.8301594853401184,],\n",
+      "[-0.2788701057434082,],\n",
+      "[-0.07654392719268799,],\n",
+      "[0.004609689582139254,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23626524209976196,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.45683401823043823,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476788520812988,],\n",
+      "[-0.12941135466098785,],\n",
+      "[0.5450205206871033,],\n",
+      "[-0.9994713068008423,],\n",
+      "[-0.028254307806491852,],\n",
+      "[0.4534774124622345,],\n",
+      "[-0.41034451127052307,],\n",
+      "[0.9649878144264221,],\n",
+      "[0.8212177157402039,],\n",
+      "[-0.9387748837471008,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23672208189964294,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.02300635538995266,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476788520812988,],\n",
+      "[-0.12941135466098785,],\n",
+      "[0.5450205206871033,],\n",
+      "[-0.9994713068008423,],\n",
+      "[-0.028254307806491852,],\n",
+      "[0.4534774124622345,],\n",
+      "[-0.41034451127052307,],\n",
+      "[0.9649878144264221,],\n",
+      "[0.8212177157402039,],\n",
+      "[-0.9387748837471008,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0020165611058473587,],\n",
+      "[-0.0007462789653800428,],\n",
+      "[-0.0035096020437777042,],\n",
+      "[0.0006717301439493895,],\n",
+      "[-0.44964367151260376,],\n",
+      "[-0.03481042757630348,],\n",
+      "[-0.6888612508773804,],\n",
+      "[-0.27779147028923035,],\n",
+      "[-0.07610014826059341,],\n",
+      "[0.004607441369444132,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23672208189964294,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3033568859100342,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476768255233765,],\n",
+      "[-0.12941060960292816,],\n",
+      "[0.545024037361145,],\n",
+      "[-0.9994719624519348,],\n",
+      "[-0.02780466340482235,],\n",
+      "[0.4535122215747833,],\n",
+      "[-0.4096556603908539,],\n",
+      "[0.9652656316757202,],\n",
+      "[0.821293830871582,],\n",
+      "[-0.9387794733047485,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2364187240600586,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20226921141147614,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476768255233765,],\n",
+      "[-0.12941060960292816,],\n",
+      "[0.545024037361145,],\n",
+      "[-0.9994719624519348,],\n",
+      "[-0.02780466340482235,],\n",
+      "[0.4535122215747833,],\n",
+      "[-0.4096556603908539,],\n",
+      "[0.9652656316757202,],\n",
+      "[0.821293830871582,],\n",
+      "[-0.9387794733047485,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.004411256406456232,],\n",
+      "[-0.001678240718320012,],\n",
+      "[-0.008352251723408699,],\n",
+      "[0.001639406313188374,],\n",
+      "[-0.4059556722640991,],\n",
+      "[-0.02966315671801567,],\n",
+      "[-0.577657163143158,],\n",
+      "[-0.27676287293434143,],\n",
+      "[-0.07593133300542831,],\n",
+      "[0.004606922622770071,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2364187240600586,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8994876146316528,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476724147796631,],\n",
+      "[-0.12940892577171326,],\n",
+      "[0.5450323820114136,],\n",
+      "[-0.9994736313819885,],\n",
+      "[-0.02739870734512806,],\n",
+      "[0.4535418748855591,],\n",
+      "[-0.4090780019760132,],\n",
+      "[0.9655423760414124,],\n",
+      "[0.8213697671890259,],\n",
+      "[-0.9387840628623962,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23551923036575317,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.23037773370742798,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476724147796631,],\n",
+      "[-0.12940892577171326,],\n",
+      "[0.5450323820114136,],\n",
+      "[-0.9994736313819885,],\n",
+      "[-0.02739870734512806,],\n",
+      "[0.4535418748855591,],\n",
+      "[-0.4090780019760132,],\n",
+      "[0.9655423760414124,],\n",
+      "[0.8213697671890259,],\n",
+      "[-0.9387840628623962,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0034689903259277344,],\n",
+      "[-0.0013594598276540637,],\n",
+      "[-0.007149753626435995,],\n",
+      "[0.0014390223659574986,],\n",
+      "[-0.40458038449287415,],\n",
+      "[-0.029037784785032272,],\n",
+      "[-0.5660421848297119,],\n",
+      "[-0.2760586738586426,],\n",
+      "[-0.07599800080060959,],\n",
+      "[0.004607519134879112,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23551923036575317,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.959955096244812,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476689577102661,],\n",
+      "[-0.12940756976604462,],\n",
+      "[0.5450395345687866,],\n",
+      "[-0.9994750618934631,],\n",
+      "[-0.02699412778019905,],\n",
+      "[0.4535709023475647,],\n",
+      "[-0.4085119664669037,],\n",
+      "[0.9658184051513672,],\n",
+      "[0.8214457631111145,],\n",
+      "[-0.938788652420044,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23455926775932312,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04846043884754181,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476689577102661,],\n",
+      "[-0.12940756976604462,],\n",
+      "[0.5450395345687866,],\n",
+      "[-0.9994750618934631,],\n",
+      "[-0.02699412778019905,],\n",
+      "[0.4535709023475647,],\n",
+      "[-0.4085119664669037,],\n",
+      "[0.9658184051513672,],\n",
+      "[0.8214457631111145,],\n",
+      "[-0.938788652420044,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0011714911088347435,],\n",
+      "[-0.00047303110477514565,],\n",
+      "[-0.0026286649517714977,],\n",
+      "[0.0005425417912192643,],\n",
+      "[-0.44303828477859497,],\n",
+      "[-0.03313018009066582,],\n",
+      "[-0.6623744964599609,],\n",
+      "[-0.2757790982723236,],\n",
+      "[-0.07615650445222855,],\n",
+      "[0.004608369432389736,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23455926775932312,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4402748942375183,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476677656173706,],\n",
+      "[-0.1294070929288864,],\n",
+      "[0.5450421571731567,],\n",
+      "[-0.9994755983352661,],\n",
+      "[-0.02655109018087387,],\n",
+      "[0.4536040425300598,],\n",
+      "[-0.40784958004951477,],\n",
+      "[0.9660941958427429,],\n",
+      "[0.8215219378471375,],\n",
+      "[-0.9387932419776917,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2341189980506897,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03064410388469696,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476677656173706,],\n",
+      "[-0.1294070929288864,],\n",
+      "[0.5450421571731567,],\n",
+      "[-0.9994755983352661,],\n",
+      "[-0.02655109018087387,],\n",
+      "[0.4536040425300598,],\n",
+      "[-0.40784958004951477,],\n",
+      "[0.9660941958427429,],\n",
+      "[0.8215219378471375,],\n",
+      "[-0.9387932419776917,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0006856402033008635,],\n",
+      "[0.0002848249569069594,],\n",
+      "[0.001674474566243589,],\n",
+      "[-0.0003545203653629869,],\n",
+      "[-0.49552106857299805,],\n",
+      "[-0.039310943335294724,],\n",
+      "[-0.8091991543769836,],\n",
+      "[-0.2759779989719391,],\n",
+      "[-0.07628136873245239,],\n",
+      "[0.004608850926160812,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2341189980506897,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.350108802318573,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476684808731079,],\n",
+      "[-0.1294073760509491,],\n",
+      "[0.545040488243103,],\n",
+      "[-0.9994752407073975,],\n",
+      "[-0.02605556882917881,],\n",
+      "[0.4536433517932892,],\n",
+      "[-0.40704038739204407,],\n",
+      "[0.966370165348053,],\n",
+      "[0.82159823179245,],\n",
+      "[-0.9387978315353394,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23446910083293915,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2189694046974182,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476684808731079,],\n",
+      "[-0.1294073760509491,],\n",
+      "[0.545040488243103,],\n",
+      "[-0.9994752407073975,],\n",
+      "[-0.02605556882917881,],\n",
+      "[0.4536433517932892,],\n",
+      "[-0.40704038739204407,],\n",
+      "[0.966370165348053,],\n",
+      "[0.82159823179245,],\n",
+      "[-0.9387978315353394,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0013485744129866362,],\n",
+      "[0.0005750178825110197,],\n",
+      "[0.003582798410207033,],\n",
+      "[-0.000778515706770122,],\n",
+      "[-0.5306122899055481,],\n",
+      "[-0.04391457140445709,],\n",
+      "[-0.9183206558227539,],\n",
+      "[-0.27660736441612244,],\n",
+      "[-0.07632444053888321,],\n",
+      "[0.004608956631273031,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23446910083293915,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9358826875686646,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476698517799377,],\n",
+      "[-0.12940795719623566,],\n",
+      "[0.5450369119644165,],\n",
+      "[-0.9994744658470154,],\n",
+      "[-0.025524957105517387,],\n",
+      "[0.45368728041648865,],\n",
+      "[-0.4061220586299896,],\n",
+      "[0.9666467905044556,],\n",
+      "[0.8216745853424072,],\n",
+      "[-0.9388024210929871,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23540498316287994,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.23267239332199097,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476698517799377,],\n",
+      "[-0.12940795719623566,],\n",
+      "[0.5450369119644165,],\n",
+      "[-0.9994744658470154,],\n",
+      "[-0.025524957105517387,],\n",
+      "[0.45368728041648865,],\n",
+      "[-0.4061220586299896,],\n",
+      "[0.9666467905044556,],\n",
+      "[0.8216745853424072,],\n",
+      "[-0.9388024210929871,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0010226140730082989,],\n",
+      "[0.0004467872786335647,],\n",
+      "[0.0029542266856878996,],\n",
+      "[-0.0006590912817046046,],\n",
+      "[-0.529813289642334,],\n",
+      "[-0.04431484639644623,],\n",
+      "[-0.924102246761322,],\n",
+      "[-0.27744895219802856,],\n",
+      "[-0.07630325108766556,],\n",
+      "[0.0046088011004030704,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23540498316287994,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9647237658500671,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476708650588989,],\n",
+      "[-0.12940840423107147,],\n",
+      "[0.5450339317321777,],\n",
+      "[-0.9994738101959229,],\n",
+      "[-0.024995144456624985,],\n",
+      "[0.45373159646987915,],\n",
+      "[-0.40519794821739197,],\n",
+      "[0.966924250125885,],\n",
+      "[0.8217508792877197,],\n",
+      "[-0.9388070106506348,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23636971414089203,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.044439055025577545,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476708650588989,],\n",
+      "[-0.12940840423107147,],\n",
+      "[0.5450339317321777,],\n",
+      "[-0.9994738101959229,],\n",
+      "[-0.024995144456624985,],\n",
+      "[0.45373159646987915,],\n",
+      "[-0.40519794821739197,],\n",
+      "[0.966924250125885,],\n",
+      "[0.8217508792877197,],\n",
+      "[-0.9388070106506348,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0003286793071310967,],\n",
+      "[0.00014713442942593247,],\n",
+      "[0.001032424159348011,],\n",
+      "[-0.00023650754883419722,],\n",
+      "[-0.4968641400337219,],\n",
+      "[-0.04039515554904938,],\n",
+      "[-0.8235366344451904,],\n",
+      "[-0.278128981590271,],\n",
+      "[-0.07625866681337357,],\n",
+      "[0.004608617164194584,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23636971414089203,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4216111898422241,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476712226867676,],\n",
+      "[-0.1294085532426834,],\n",
+      "[0.5450329184532166,],\n",
+      "[-0.9994735717773438,],\n",
+      "[-0.02449828013777733,],\n",
+      "[0.4537719786167145,],\n",
+      "[-0.4043744206428528,],\n",
+      "[0.967202365398407,],\n",
+      "[0.8218271136283875,],\n",
+      "[-0.9388116002082825,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23679132759571075,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03270099684596062,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476712226867676,],\n",
+      "[-0.1294085532426834,],\n",
+      "[0.5450329184532166,],\n",
+      "[-0.9994735717773438,],\n",
+      "[-0.02449828013777733,],\n",
+      "[0.4537719786167145,],\n",
+      "[-0.4043744206428528,],\n",
+      "[0.967202365398407,],\n",
+      "[0.8218271136283875,],\n",
+      "[-0.9388116002082825,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00020731134281959385,],\n",
+      "[-9.523936751065776e-05,],\n",
+      "[-0.0007082708762027323,],\n",
+      "[0.00016654179489705712,],\n",
+      "[-0.45400670170783997,],\n",
+      "[-0.034679509699344635,],\n",
+      "[-0.678383469581604,],\n",
+      "[-0.27829375863075256,],\n",
+      "[-0.0762254148721695,],\n",
+      "[0.004608510993421078,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23679132759571075,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3616682291030884,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476710438728333,],\n",
+      "[-0.12940846383571625,],\n",
+      "[0.5450336337089539,],\n",
+      "[-0.9994737505912781,],\n",
+      "[-0.024044273421168327,],\n",
+      "[0.45380666851997375,],\n",
+      "[-0.4036960303783417,],\n",
+      "[0.9674806594848633,],\n",
+      "[0.8219033479690552,],\n",
+      "[-0.9388161897659302,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23642966151237488,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20604148507118225,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476710438728333,],\n",
+      "[-0.12940846383571625,],\n",
+      "[0.5450336337089539,],\n",
+      "[-0.9994737505912781,],\n",
+      "[-0.024044273421168327,],\n",
+      "[0.45380666851997375,],\n",
+      "[-0.4036960303783417,],\n",
+      "[0.9674806594848633,],\n",
+      "[0.8219033479690552,],\n",
+      "[-0.9388161897659302,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00038255733670666814,],\n",
+      "[-0.00018077599816024303,],\n",
+      "[-0.0014222662430256605,],\n",
+      "[0.0003430885844863951,],\n",
+      "[-0.4274716377258301,],\n",
+      "[-0.030751854181289673,],\n",
+      "[-0.5771302580833435,],\n",
+      "[-0.2778281569480896,],\n",
+      "[-0.07621481269598007,],\n",
+      "[0.004608518909662962,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23642966151237488,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9078367948532104,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476706862449646,],\n",
+      "[-0.12940828502178192,],\n",
+      "[0.5450350642204285,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.02361680194735527,],\n",
+      "[0.45383742451667786,],\n",
+      "[-0.40311890840530396,],\n",
+      "[0.9677584767341614,],\n",
+      "[0.8219795823097229,],\n",
+      "[-0.9388207793235779,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2355218231678009,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.19534988701343536,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476706862449646,],\n",
+      "[-0.12940828502178192,],\n",
+      "[0.5450350642204285,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.02361680194735527,],\n",
+      "[0.45383742451667786,],\n",
+      "[-0.40311890840530396,],\n",
+      "[0.9677584767341614,],\n",
+      "[0.8219795823097229,],\n",
+      "[-0.9388207793235779,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00027381142717786133,],\n",
+      "[-0.00013330281944945455,],\n",
+      "[-0.0011081506963819265,],\n",
+      "[0.00027412609779275954,],\n",
+      "[-0.43114352226257324,],\n",
+      "[-0.030899502336978912,],\n",
+      "[-0.5811771154403687,],\n",
+      "[-0.27696874737739563,],\n",
+      "[-0.07622121274471283,],\n",
+      "[0.004608532413840294,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2355218231678009,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8839670419692993,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476703882217407,],\n",
+      "[-0.12940815091133118,],\n",
+      "[0.5450361967086792,],\n",
+      "[-0.9994744062423706,],\n",
+      "[-0.02318565919995308,],\n",
+      "[0.4538683295249939,],\n",
+      "[-0.4025377333164215,],\n",
+      "[0.9680354595184326,],\n",
+      "[0.8220558166503906,],\n",
+      "[-0.9388253688812256,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23463785648345947,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.023073649033904076,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476703882217407,],\n",
+      "[-0.12940815091133118,],\n",
+      "[0.5450361967086792,],\n",
+      "[-0.9994744062423706,],\n",
+      "[-0.02318565919995308,],\n",
+      "[0.4538683295249939,],\n",
+      "[-0.4025377333164215,],\n",
+      "[0.9680354595184326,],\n",
+      "[0.8220558166503906,],\n",
+      "[-0.9388253688812256,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[6.916785059729591e-05,],\n",
+      "[-3.46895722032059e-05,],\n",
+      "[-0.0003047133213840425,],\n",
+      "[7.72983767092228e-05,],\n",
+      "[-0.45944806933403015,],\n",
+      "[-0.03476162999868393,],\n",
+      "[-0.6881023049354553,],\n",
+      "[-0.2762015759944916,],\n",
+      "[-0.07623366266489029,],\n",
+      "[0.0046085878275334835,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23463785648345947,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3038000762462616,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.647670328617096,],\n",
+      "[-0.1294081211090088,],\n",
+      "[0.5450364947319031,],\n",
+      "[-0.9994744658470154,],\n",
+      "[-0.022726211696863174,],\n",
+      "[0.45390307903289795,],\n",
+      "[-0.401849627494812,],\n",
+      "[0.9683116674423218,],\n",
+      "[0.8221320509910583,],\n",
+      "[-0.9388299584388733,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2343340516090393,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.05770714581012726,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.647670328617096,],\n",
+      "[-0.1294081211090088,],\n",
+      "[0.5450364947319031,],\n",
+      "[-0.9994744658470154,],\n",
+      "[-0.022726211696863174,],\n",
+      "[0.45390307903289795,],\n",
+      "[-0.401849627494812,],\n",
+      "[0.9683116674423218,],\n",
+      "[0.8221320509910583,],\n",
+      "[-0.9388299584388733,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.039934618864208e-05,],\n",
+      "[4.146833089180291e-05,],\n",
+      "[0.00038536760257557034,],\n",
+      "[-0.00010029251279775053,],\n",
+      "[-0.49322372674942017,],\n",
+      "[-0.03982311487197876,],\n",
+      "[-0.8328792452812195,],\n",
+      "[-0.27599626779556274,],\n",
+      "[-0.07624180614948273,],\n",
+      "[0.00460859015583992,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2343340516090393,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4804461896419525,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476703882217407,],\n",
+      "[-0.12940816581249237,],\n",
+      "[0.5450361371040344,],\n",
+      "[-0.9994743466377258,],\n",
+      "[-0.022232986986637115,],\n",
+      "[0.4539428949356079,],\n",
+      "[-0.4010167419910431,],\n",
+      "[0.9685876369476318,],\n",
+      "[0.8222082853317261,],\n",
+      "[-0.938834547996521,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23481449484825134,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.24598120152950287,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476703882217407,],\n",
+      "[-0.12940816581249237,],\n",
+      "[0.5450361371040344,],\n",
+      "[-0.9994743466377258,],\n",
+      "[-0.022232986986637115,],\n",
+      "[0.4539428949356079,],\n",
+      "[-0.4010167419910431,],\n",
+      "[0.9685876369476318,],\n",
+      "[0.8222082853317261,],\n",
+      "[-0.938834547996521,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00012200402852613479,],\n",
+      "[6.456828123191372e-05,],\n",
+      "[0.0006358509417623281,],\n",
+      "[-0.00016987326671369374,],\n",
+      "[-0.5125383734703064,],\n",
+      "[-0.043064557015895844,],\n",
+      "[-0.9271420240402222,],\n",
+      "[-0.2765122652053833,],\n",
+      "[-0.07624410092830658,],\n",
+      "[0.0046085878275334835,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23481449484825134,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9919288754463196,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940822541713715,],\n",
+      "[0.5450354814529419,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.02172044850885868,],\n",
+      "[0.45398595929145813,],\n",
+      "[-0.4000895917415619,],\n",
+      "[0.9688641428947449,],\n",
+      "[0.8222845196723938,],\n",
+      "[-0.9388391375541687,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2358064204454422,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.21113260090351105,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940822541713715,],\n",
+      "[0.5450354814529419,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.02172044850885868,],\n",
+      "[0.45398595929145813,],\n",
+      "[-0.4000895917415619,],\n",
+      "[0.9688641428947449,],\n",
+      "[0.8222845196723938,],\n",
+      "[-0.9388391375541687,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.307662210427225e-05,],\n",
+      "[4.504591561271809e-05,],\n",
+      "[0.00047057404299266636,],\n",
+      "[-0.00012910763325635344,],\n",
+      "[-0.5078024864196777,],\n",
+      "[-0.042677439749240875,],\n",
+      "[-0.9134360551834106,],\n",
+      "[-0.2774830162525177,],\n",
+      "[-0.07624195516109467,],\n",
+      "[0.004608584567904472,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2358064204454422,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9189839959144592,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.647670567035675,],\n",
+      "[-0.12940827012062073,],\n",
+      "[0.5450350046157837,],\n",
+      "[-0.999474048614502,],\n",
+      "[-0.02121264673769474,],\n",
+      "[0.4540286362171173,],\n",
+      "[-0.39917615056037903,],\n",
+      "[0.9691416025161743,],\n",
+      "[0.8223607540130615,],\n",
+      "[-0.9388437271118164,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23672540485858917,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.023211318999528885,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.647670567035675,],\n",
+      "[-0.12940827012062073,],\n",
+      "[0.5450350046157837,],\n",
+      "[-0.999474048614502,],\n",
+      "[-0.02121264673769474,],\n",
+      "[0.4540286362171173,],\n",
+      "[-0.39917615056037903,],\n",
+      "[0.9691416025161743,],\n",
+      "[0.8223607540130615,],\n",
+      "[-0.9388437271118164,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.0244993720552884e-05,],\n",
+      "[1.124908339988906e-05,],\n",
+      "[0.00012463661551009864,],\n",
+      "[-3.5115594073431566e-05,],\n",
+      "[-0.4840945303440094,],\n",
+      "[-0.03905361145734787,],\n",
+      "[-0.8009780049324036,],\n",
+      "[-0.2783510982990265,],\n",
+      "[-0.07623828947544098,],\n",
+      "[0.004608582705259323,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23672540485858917,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3047049939632416,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.647670567035675,],\n",
+      "[-0.12940828502178192,],\n",
+      "[0.5450348854064941,],\n",
+      "[-0.9994739890098572,],\n",
+      "[-0.02072855271399021,],\n",
+      "[0.4540676772594452,],\n",
+      "[-0.39837518334388733,],\n",
+      "[0.9694199562072754,],\n",
+      "[0.8224369883537292,],\n",
+      "[-0.9388483166694641,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23703010380268097,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.05737204849720001,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.647670567035675,],\n",
+      "[-0.12940828502178192,],\n",
+      "[0.5450348854064941,],\n",
+      "[-0.9994739890098572,],\n",
+      "[-0.02072855271399021,],\n",
+      "[0.4540676772594452,],\n",
+      "[-0.39837518334388733,],\n",
+      "[0.9694199562072754,],\n",
+      "[0.8224369883537292,],\n",
+      "[-0.9388483166694641,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.3392374714603648e-05,],\n",
+      "[-1.334346325165825e-05,],\n",
+      "[-0.00015659737982787192,],\n",
+      "[4.528449790086597e-05,],\n",
+      "[-0.45710378885269165,],\n",
+      "[-0.034528471529483795,],\n",
+      "[-0.6587396264076233,],\n",
+      "[-0.27861401438713074,],\n",
+      "[-0.0762360692024231,],\n",
+      "[0.004608582239598036,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23703010380268097,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.47904863953590393,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.647670567035675,],\n",
+      "[-0.12940827012062073,],\n",
+      "[0.5450350642204285,],\n",
+      "[-0.999474048614502,],\n",
+      "[-0.02027144841849804,],\n",
+      "[0.45410221815109253,],\n",
+      "[-0.3977164328098297,],\n",
+      "[0.9696985483169556,],\n",
+      "[0.822513222694397,],\n",
+      "[-0.9388529062271118,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23655106127262115,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22917543351650238,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.647670567035675,],\n",
+      "[-0.12940827012062073,],\n",
+      "[0.5450350642204285,],\n",
+      "[-0.999474048614502,],\n",
+      "[-0.02027144841849804,],\n",
+      "[0.45410221815109253,],\n",
+      "[-0.3977164328098297,],\n",
+      "[0.9696985483169556,],\n",
+      "[0.822513222694397,],\n",
+      "[-0.9388529062271118,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.436034239712171e-05,],\n",
+      "[-2.0166758986306377e-05,],\n",
+      "[-0.00025027518859133124,],\n",
+      "[7.423558417940512e-05,],\n",
+      "[-0.4429524838924408,],\n",
+      "[-0.031876254826784134,],\n",
+      "[-0.571914792060852,],\n",
+      "[-0.2781231105327606,],\n",
+      "[-0.0762360543012619,],\n",
+      "[0.004608582705259323,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23655106127262115,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9574460387229919,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940825521945953,],\n",
+      "[0.5450353026390076,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.019828496500849724,],\n",
+      "[0.45413410663604736,],\n",
+      "[-0.39714452624320984,],\n",
+      "[0.9699766635894775,],\n",
+      "[0.8225894570350647,],\n",
+      "[-0.9388574957847595,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23559361696243286,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.17684106528759003,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940825521945953,],\n",
+      "[0.5450353026390076,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.019828496500849724,],\n",
+      "[0.45413410663604736,],\n",
+      "[-0.39714452624320984,],\n",
+      "[0.9699766635894775,],\n",
+      "[0.8225894570350647,],\n",
+      "[-0.9388574957847595,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.2182408429216594e-05,],\n",
+      "[-1.3414494787866715e-05,],\n",
+      "[-0.00017586754984222353,],\n",
+      "[5.34851205884479e-05,],\n",
+      "[-0.44829273223876953,],\n",
+      "[-0.03253724426031113,],\n",
+      "[-0.592217743396759,],\n",
+      "[-0.27717772126197815,],\n",
+      "[-0.07623603194952011,],\n",
+      "[0.00460858317092061,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23559361696243286,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8410499691963196,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354814529419,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.019380204379558563,],\n",
+      "[0.4541666507720947,],\n",
+      "[-0.39655229449272156,],\n",
+      "[0.9702538251876831,],\n",
+      "[0.8226656913757324,],\n",
+      "[-0.9388620853424072,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2347525656223297,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.010194552130997181,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354814529419,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.019380204379558563,],\n",
+      "[0.4541666507720947,],\n",
+      "[-0.39655229449272156,],\n",
+      "[0.9702538251876831,],\n",
+      "[0.8226656913757324,],\n",
+      "[-0.9388620853424072,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.914213266398292e-06,],\n",
+      "[-2.43805993704882e-06,],\n",
+      "[-3.377476241439581e-05,],\n",
+      "[1.0532463420531712e-05,],\n",
+      "[-0.4675006866455078,],\n",
+      "[-0.03584767505526543,],\n",
+      "[-0.7076250314712524,],\n",
+      "[-0.2763546407222748,],\n",
+      "[-0.0762375071644783,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2347525656223297,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.20193609595298767,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450355410575867,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.018912702798843384,],\n",
+      "[0.4542025029659271,],\n",
+      "[-0.3958446681499481,],\n",
+      "[0.9705301523208618,],\n",
+      "[0.8227419257164001,],\n",
+      "[-0.9388666749000549,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2345506250858307,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08114539831876755,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450355410575867,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.018912702798843384,],\n",
+      "[0.4542025029659271,],\n",
+      "[-0.3958446681499481,],\n",
+      "[0.9705301523208618,],\n",
+      "[0.8227419257164001,],\n",
+      "[-0.9388666749000549,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.11587324278662e-06,],\n",
+      "[5.197157861402957e-06,],\n",
+      "[7.617608935106546e-05,],\n",
+      "[-2.4371242034249008e-05,],\n",
+      "[-0.48779839277267456,],\n",
+      "[-0.03965812548995018,],\n",
+      "[-0.8474541902542114,],\n",
+      "[-0.27617061138153076,],\n",
+      "[-0.076238252222538,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2345506250858307,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5697211623191833,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354814529419,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.018424903973937035,],\n",
+      "[0.45424216985702515,],\n",
+      "[-0.3949972093105316,],\n",
+      "[0.9708063006401062,],\n",
+      "[0.8228181600570679,],\n",
+      "[-0.9388712644577026,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23512034118175507,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.251382976770401,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354814529419,],\n",
+      "[-0.9994741678237915,],\n",
+      "[-0.018424903973937035,],\n",
+      "[0.45424216985702515,],\n",
+      "[-0.3949972093105316,],\n",
+      "[0.9708063006401062,],\n",
+      "[0.8228181600570679,],\n",
+      "[-0.9388712644577026,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.049817365128547e-05,],\n",
+      "[6.8961189754190855e-06,],\n",
+      "[0.00010711442155297846,],\n",
+      "[-3.5181281418772414e-05,],\n",
+      "[-0.4974963665008545,],\n",
+      "[-0.04171164706349373,],\n",
+      "[-0.9249905943870544,],\n",
+      "[-0.2767432630062103,],\n",
+      "[-0.07623825967311859,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23512034118175507,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0027635097503662,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353622436523,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.01792740821838379,],\n",
+      "[0.4542838931083679,],\n",
+      "[-0.39407220482826233,],\n",
+      "[0.9710830450057983,],\n",
+      "[0.8228943943977356,],\n",
+      "[-0.9388758540153503,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23612309992313385,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.17292024195194244,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353622436523,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.01792740821838379,],\n",
+      "[0.4542838931083679,],\n",
+      "[-0.39407220482826233,],\n",
+      "[0.9710830450057983,],\n",
+      "[0.8228943943977356,],\n",
+      "[-0.9388758540153503,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.398974164767424e-06,],\n",
+      "[4.306416940380586e-06,],\n",
+      "[7.094666216289625e-05,],\n",
+      "[-2.393097383901477e-05,],\n",
+      "[-0.4922282099723816,],\n",
+      "[-0.04091664031147957,],\n",
+      "[-0.8929793834686279,],\n",
+      "[-0.2777400016784668,],\n",
+      "[-0.0762375220656395,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23612309992313385,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.8316741585731506,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353026390076,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.01743518002331257,],\n",
+      "[0.4543248116970062,],\n",
+      "[-0.3931792378425598,],\n",
+      "[0.9713608026504517,],\n",
+      "[0.8229706287384033,],\n",
+      "[-0.938880443572998,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23695477843284607,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.006238914560526609,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353026390076,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.01743518002331257,],\n",
+      "[0.4543248116970062,],\n",
+      "[-0.3931792378425598,],\n",
+      "[0.9713608026504517,],\n",
+      "[0.8229706287384033,],\n",
+      "[-0.938880443572998,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.932703394748387e-07,],\n",
+      "[6.161905048429617e-07,],\n",
+      "[1.0763421414594632e-05,],\n",
+      "[-3.728023557414417e-06,],\n",
+      "[-0.476603627204895,],\n",
+      "[-0.03791322559118271,],\n",
+      "[-0.7733328342437744,],\n",
+      "[-0.2785656750202179,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23695477843284607,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.15797361731529236,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353026390076,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.016958575695753098,],\n",
+      "[0.4543627202510834,],\n",
+      "[-0.3924058973789215,],\n",
+      "[0.9716393947601318,],\n",
+      "[0.823046863079071,],\n",
+      "[-0.9388850331306458,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23711274564266205,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09352079033851624,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353026390076,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.016958575695753098,],\n",
+      "[0.4543627202510834,],\n",
+      "[-0.3924058973789215,],\n",
+      "[0.9716393947601318,],\n",
+      "[0.823046863079071,],\n",
+      "[-0.9388850331306458,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.541682079026941e-06,],\n",
+      "[-1.800654104044952e-06,],\n",
+      "[-3.330299296067096e-05,],\n",
+      "[1.1837521014967933e-05,],\n",
+      "[-0.4608129858970642,],\n",
+      "[-0.03462665528059006,],\n",
+      "[-0.6391595602035522,],\n",
+      "[-0.27871471643447876,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23711274564266205,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6116234064102173,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353622436523,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.016497762873768806,],\n",
+      "[0.4543973505496979,],\n",
+      "[-0.3917667269706726,],\n",
+      "[0.9719181060791016,],\n",
+      "[0.8231230974197388,],\n",
+      "[-0.9388896226882935,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23650112748146057,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.25600048899650574,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450353622436523,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.016497762873768806,],\n",
+      "[0.4543973505496979,],\n",
+      "[-0.3917667269706726,],\n",
+      "[0.9719181060791016,],\n",
+      "[0.8231230974197388,],\n",
+      "[-0.9388896226882935,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.0904805043974193e-06,],\n",
+      "[-2.253634420412709e-06,],\n",
+      "[-4.406213338370435e-05,],\n",
+      "[1.6062123904703185e-05,],\n",
+      "[-0.4540502727031708,],\n",
+      "[-0.03302532061934471,],\n",
+      "[-0.5687519311904907,],\n",
+      "[-0.2781031131744385,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23650112748146057,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0119307041168213,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.016043713316321373,],\n",
+      "[0.45443037152290344,],\n",
+      "[-0.39119797945022583,],\n",
+      "[0.9721962213516235,],\n",
+      "[0.8231993317604065,],\n",
+      "[-0.9388942122459412,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23548918962478638,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.16062183678150177,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.016043713316321373,],\n",
+      "[0.45443037152290344,],\n",
+      "[-0.39119797945022583,],\n",
+      "[0.9721962213516235,],\n",
+      "[0.8231993317604065,],\n",
+      "[-0.9388942122459412,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.799062602003687e-06,],\n",
+      "[-1.3519287449526018e-06,],\n",
+      "[-2.791980114125181e-05,],\n",
+      "[1.0434179785079323e-05,],\n",
+      "[-0.45902979373931885,],\n",
+      "[-0.033904824405908585,],\n",
+      "[-0.6038573980331421,],\n",
+      "[-0.27709418535232544,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23548918962478638,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8015528321266174,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.015584683045744896,],\n",
+      "[0.45446428656578064,],\n",
+      "[-0.39059412479400635,],\n",
+      "[0.9724733233451843,],\n",
+      "[0.8232755661010742,],\n",
+      "[-0.9388988018035889,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23468764126300812,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0030416524969041348,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.015584683045744896,],\n",
+      "[0.45446428656578064,],\n",
+      "[-0.39059412479400635,],\n",
+      "[0.9724733233451843,],\n",
+      "[0.8232755661010742,],\n",
+      "[-0.9388988018035889,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.8194344875155366e-07,],\n",
+      "[-1.4080768551139045e-07,],\n",
+      "[-3.072902700296254e-06,],\n",
+      "[1.177560420728696e-06,],\n",
+      "[-0.47129225730895996,],\n",
+      "[-0.036520715802907944,],\n",
+      "[-0.7247166633605957,],\n",
+      "[-0.27629563212394714,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23468764126300812,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.11030227690935135,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.015113390982151031,],\n",
+      "[0.45450079441070557,],\n",
+      "[-0.38986942172050476,],\n",
+      "[0.9727495908737183,],\n",
+      "[0.8233518004417419,],\n",
+      "[-0.9389033913612366,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23457734286785126,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1020885780453682,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.015113390982151031,],\n",
+      "[0.45450079441070557,],\n",
+      "[-0.38986942172050476,],\n",
+      "[0.9727495908737183,],\n",
+      "[0.8233518004417419,],\n",
+      "[-0.9389033913612366,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.746565415800433e-07,],\n",
+      "[6.161866394904791e-07,],\n",
+      "[1.4230129636416677e-05,],\n",
+      "[-5.594848516921047e-06,],\n",
+      "[-0.4828784167766571,],\n",
+      "[-0.03919300064444542,],\n",
+      "[-0.8566137552261353,],\n",
+      "[-0.27618831396102905,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23457734286785126,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6390267610549927,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.014630512334406376,],\n",
+      "[0.4545399844646454,],\n",
+      "[-0.38901281356811523,],\n",
+      "[0.9730257987976074,],\n",
+      "[0.8234280347824097,],\n",
+      "[-0.9389079809188843,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23521636426448822,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.24412107467651367,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.014630512334406376,],\n",
+      "[0.4545399844646454,],\n",
+      "[-0.38901281356811523,],\n",
+      "[0.9730257987976074,],\n",
+      "[0.8234280347824097,],\n",
+      "[-0.9389079809188843,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.80362506450183e-07,],\n",
+      "[7.18187720849528e-07,],\n",
+      "[1.7577907783561386e-05,],\n",
+      "[-7.095343789842445e-06,],\n",
+      "[-0.4872986376285553,],\n",
+      "[-0.0403621681034565,],\n",
+      "[-0.9160696268081665,],\n",
+      "[-0.276827335357666,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23521636426448822,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9881730675697327,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.014143213629722595,],\n",
+      "[0.4545803368091583,],\n",
+      "[-0.38809674978256226,],\n",
+      "[0.9733026027679443,],\n",
+      "[0.8235042691230774,],\n",
+      "[-0.938912570476532,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23620453476905823,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13052301108837128,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.014143213629722595,],\n",
+      "[0.4545803368091583,],\n",
+      "[-0.38809674978256226,],\n",
+      "[0.9733026027679443,],\n",
+      "[0.8235042691230774,],\n",
+      "[-0.938912570476532,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-4.7308716943916806e-07,],\n",
+      "[3.954105238790362e-07,],\n",
+      "[1.0264103366353083e-05,],\n",
+      "[-4.2548599594738334e-06,],\n",
+      "[-0.48288607597351074,],\n",
+      "[-0.03945842385292053,],\n",
+      "[-0.8678399324417114,],\n",
+      "[-0.27781251072883606,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23620453476905823,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7225592732429504,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.013660327531397343,],\n",
+      "[0.45461979508399963,],\n",
+      "[-0.38722890615463257,],\n",
+      "[0.9735804200172424,],\n",
+      "[0.8235805034637451,],\n",
+      "[-0.9389171600341797,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2369270920753479,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([2.7471611474538804e-07,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.013660327531397343,],\n",
+      "[0.45461979508399963,],\n",
+      "[-0.38722890615463257,],\n",
+      "[0.9735804200172424,],\n",
+      "[0.8235805034637451,],\n",
+      "[-0.9389171600341797,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.04402020151673e-10,],\n",
+      "[4.3226255908024314e-10,],\n",
+      "[1.1894067064588398e-08,],\n",
+      "[-5.062353647389273e-09,],\n",
+      "[-0.47317981719970703,],\n",
+      "[-0.03716330975294113,],\n",
+      "[-0.7457853555679321,],\n",
+      "[-0.27853506803512573,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2369270920753479,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.0010482671204954386,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.013187147676944733,],\n",
+      "[0.4546569585800171,],\n",
+      "[-0.3864831328392029,],\n",
+      "[0.9738589525222778,],\n",
+      "[0.8236567378044128,],\n",
+      "[-0.9389217495918274,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23692813515663147,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1373223513364792,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.013187147676944733,],\n",
+      "[0.4546569585800171,],\n",
+      "[-0.3864831328392029,],\n",
+      "[0.9738589525222778,],\n",
+      "[0.8236567378044128,],\n",
+      "[-0.9389217495918274,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.620860755087051e-07,],\n",
+      "[-2.3077440403085347e-07,],\n",
+      "[-6.721117642882746e-06,],\n",
+      "[2.935278871518676e-06,],\n",
+      "[-0.46442094445228577,],\n",
+      "[-0.03493456542491913,],\n",
+      "[-0.6232568025588989,],\n",
+      "[-0.2785361111164093,],\n",
+      "[-0.07623752951622009,],\n",
+      "[0.004608583636581898,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23692813515663147,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7411404848098755,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.012722726911306381,],\n",
+      "[0.45469188690185547,],\n",
+      "[-0.385859876871109,],\n",
+      "[0.9741374850273132,],\n",
+      "[0.8237329721450806,],\n",
+      "[-0.9389263391494751,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23618699610233307,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [5/10], Loss: 0.1373\n",
+      "loss_antes tensor([0.6579055786132812,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6476705074310303,],\n",
+      "[-0.12940824031829834,],\n",
+      "[0.5450354218482971,],\n",
+      "[-0.9994741082191467,],\n",
+      "[-0.012722726911306381,],\n",
+      "[0.45469188690185547,],\n",
+      "[-0.385859876871109,],\n",
+      "[0.9741374850273132,],\n",
+      "[0.8237329721450806,],\n",
+      "[-0.9389263391494751,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.33010154962539673,],\n",
+      "[-0.13919322192668915,],\n",
+      "[-0.4332006275653839,],\n",
+      "[0.0695735290646553,],\n",
+      "[-0.2812666893005371,],\n",
+      "[-0.022801999002695084,],\n",
+      "[-0.45390647649765015,],\n",
+      "[-0.1881387084722519,],\n",
+      "[-0.05002284049987793,],\n",
+      "[0.0030977115966379642,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23618699610233307,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.6222267150878906,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6473404169082642,],\n",
+      "[-0.1292690485715866,],\n",
+      "[0.5454686284065247,],\n",
+      "[-0.9995436668395996,],\n",
+      "[-0.01244146004319191,],\n",
+      "[0.45471468567848206,],\n",
+      "[-0.385405957698822,],\n",
+      "[0.9743255972862244,],\n",
+      "[0.8237829804420471,],\n",
+      "[-0.9389294385910034,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23456476628780365,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6434317231178284,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6473404169082642,],\n",
+      "[-0.1292690485715866,],\n",
+      "[0.5454686284065247,],\n",
+      "[-0.9995436668395996,],\n",
+      "[-0.01244146004319191,],\n",
+      "[0.45471468567848206,],\n",
+      "[-0.385405957698822,],\n",
+      "[0.9743255972862244,],\n",
+      "[0.8237829804420471,],\n",
+      "[-0.9389294385910034,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.35126182436943054,],\n",
+      "[-0.1377418488264084,],\n",
+      "[-0.39933329820632935,],\n",
+      "[0.06218485161662102,],\n",
+      "[-0.28324130177497864,],\n",
+      "[-0.022183872759342194,],\n",
+      "[-0.45558643341064453,],\n",
+      "[-0.19773755967617035,],\n",
+      "[-0.04677377641201019,],\n",
+      "[0.0037098743487149477,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23456476628780365,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.6042829751968384,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6469891667366028,],\n",
+      "[-0.12913130223751068,],\n",
+      "[0.5458679795265198,],\n",
+      "[-0.9996058344841003,],\n",
+      "[-0.012158218771219254,],\n",
+      "[0.4547368586063385,],\n",
+      "[-0.3849503695964813,],\n",
+      "[0.9745233058929443,],\n",
+      "[0.8238297700881958,],\n",
+      "[-0.9389331340789795,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.23296047747135162,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.0104846954345703,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6469891667366028,],\n",
+      "[-0.12913130223751068,],\n",
+      "[0.5458679795265198,],\n",
+      "[-0.9996058344841003,],\n",
+      "[-0.012158218771219254,],\n",
+      "[0.4547368586063385,],\n",
+      "[-0.3849503695964813,],\n",
+      "[0.9745233058929443,],\n",
+      "[0.8238297700881958,],\n",
+      "[-0.9389331340789795,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.45241039991378784,],\n",
+      "[-0.16614486277103424,],\n",
+      "[-0.456428587436676,],\n",
+      "[0.06947144120931625,],\n",
+      "[-0.23601728677749634,],\n",
+      "[-0.017989493906497955,],\n",
+      "[-0.38076335191726685,],\n",
+      "[-0.18906155228614807,],\n",
+      "[-0.03713299334049225,],\n",
+      "[0.0039786286652088165,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.23296047747135162,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.01045560836792,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6465367674827576,],\n",
+      "[-0.12896515429019928,],\n",
+      "[0.5463244318962097,],\n",
+      "[-0.9996753334999084,],\n",
+      "[-0.01192220114171505,],\n",
+      "[0.45475485920906067,],\n",
+      "[-0.38456961512565613,],\n",
+      "[0.9747123718261719,],\n",
+      "[0.8238669037818909,],\n",
+      "[-0.9389371275901794,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2309500277042389,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.4477484226226807,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6465367674827576,],\n",
+      "[-0.12896515429019928,],\n",
+      "[0.5463244318962097,],\n",
+      "[-0.9996753334999084,],\n",
+      "[-0.01192220114171505,],\n",
+      "[0.45475485920906067,],\n",
+      "[-0.38456961512565613,],\n",
+      "[0.9747123718261719,],\n",
+      "[0.8238669037818909,],\n",
+      "[-0.9389371275901794,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.530709981918335,],\n",
+      "[-0.18452855944633484,],\n",
+      "[-0.48852038383483887,],\n",
+      "[0.07319517433643341,],\n",
+      "[-0.19121313095092773,],\n",
+      "[-0.013948273845016956,],\n",
+      "[-0.30741867423057556,],\n",
+      "[-0.18509718775749207,],\n",
+      "[-0.029355982318520546,],\n",
+      "[0.004348463844507933,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2309500277042389,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.40645170211792,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6460060477256775,],\n",
+      "[-0.12878061830997467,],\n",
+      "[0.5468129515647888,],\n",
+      "[-0.9997485280036926,],\n",
+      "[-0.011730987578630447,],\n",
+      "[0.4547688066959381,],\n",
+      "[-0.38426220417022705,],\n",
+      "[0.9748974442481995,],\n",
+      "[0.8238962888717651,],\n",
+      "[-0.938941478729248,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22854357957839966,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.4058836698532104,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6460060477256775,],\n",
+      "[-0.12878061830997467,],\n",
+      "[0.5468129515647888,],\n",
+      "[-0.9997485280036926,],\n",
+      "[-0.011730987578630447,],\n",
+      "[0.4547688066959381,],\n",
+      "[-0.38426220417022705,],\n",
+      "[0.9748974442481995,],\n",
+      "[0.8238962888717651,],\n",
+      "[-0.938941478729248,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4890628159046173,],\n",
+      "[-0.16311544179916382,],\n",
+      "[-0.4228256940841675,],\n",
+      "[0.06278499960899353,],\n",
+      "[-0.1974852979183197,],\n",
+      "[-0.0135266724973917,],\n",
+      "[-0.31199413537979126,],\n",
+      "[-0.19790710508823395,],\n",
+      "[-0.03166484460234642,],\n",
+      "[0.004945984575897455,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22854357957839966,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.37139892578125,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6455169916152954,],\n",
+      "[-0.12861751019954681,],\n",
+      "[0.5472357869148254,],\n",
+      "[-0.9998112916946411,],\n",
+      "[-0.011533502489328384,],\n",
+      "[0.4547823369503021,],\n",
+      "[-0.38395020365715027,],\n",
+      "[0.9750953316688538,],\n",
+      "[0.8239279389381409,],\n",
+      "[-0.9389464259147644,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22617217898368835,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7721152305603027,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6455169916152954,],\n",
+      "[-0.12861751019954681,],\n",
+      "[0.5472357869148254,],\n",
+      "[-0.9998112916946411,],\n",
+      "[-0.011533502489328384,],\n",
+      "[0.4547823369503021,],\n",
+      "[-0.38395020365715027,],\n",
+      "[0.9750953316688538,],\n",
+      "[0.8239279389381409,],\n",
+      "[-0.9389464259147644,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3243308663368225,],\n",
+      "[-0.10512252897024155,],\n",
+      "[-0.2708267271518707,],\n",
+      "[0.040122006088495255,],\n",
+      "[-0.2698360085487366,],\n",
+      "[-0.017636358737945557,],\n",
+      "[-0.42184582352638245,],\n",
+      "[-0.22399385273456573,],\n",
+      "[-0.04440409317612648,],\n",
+      "[0.005600982345640659,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22617217898368835,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.75740385055542,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6451926827430725,],\n",
+      "[-0.12851238250732422,],\n",
+      "[0.5475066304206848,],\n",
+      "[-0.999851405620575,],\n",
+      "[-0.011263666674494743,],\n",
+      "[0.45479997992515564,],\n",
+      "[-0.38352835178375244,],\n",
+      "[0.9753193259239197,],\n",
+      "[0.8239723443984985,],\n",
+      "[-0.9389520287513733,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22441478073596954,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1531953662633896,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6451926827430725,],\n",
+      "[-0.12851238250732422,],\n",
+      "[0.5475066304206848,],\n",
+      "[-0.999851405620575,],\n",
+      "[-0.011263666674494743,],\n",
+      "[0.45479997992515564,],\n",
+      "[-0.38352835178375244,],\n",
+      "[0.9753193259239197,],\n",
+      "[0.8239723443984985,],\n",
+      "[-0.9389520287513733,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.12429937720298767,],\n",
+      "[-0.039591386914253235,],\n",
+      "[-0.10280609875917435,],\n",
+      "[0.015291483141481876,],\n",
+      "[-0.3807871341705322,],\n",
+      "[-0.024401532486081123,],\n",
+      "[-0.5978355407714844,],\n",
+      "[-0.25125929713249207,],\n",
+      "[-0.06031448394060135,],\n",
+      "[0.006088242866098881,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22441478073596954,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7828041315078735,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450684070587158,],\n",
+      "[-0.12847279012203217,],\n",
+      "[0.5476094484329224,],\n",
+      "[-0.9998667240142822,],\n",
+      "[-0.010882879607379436,],\n",
+      "[0.45482438802719116,],\n",
+      "[-0.3829305171966553,],\n",
+      "[0.975570559501648,],\n",
+      "[0.8240326642990112,],\n",
+      "[-0.9389581084251404,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22363197803497314,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0023999034892767668,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450684070587158,],\n",
+      "[-0.12847279012203217,],\n",
+      "[0.5476094484329224,],\n",
+      "[-0.9998667240142822,],\n",
+      "[-0.010882879607379436,],\n",
+      "[0.45482438802719116,],\n",
+      "[-0.3829305171966553,],\n",
+      "[0.975570559501648,],\n",
+      "[0.8240326642990112,],\n",
+      "[-0.9389581084251404,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.01294650137424469,],\n",
+      "[0.004087480250746012,],\n",
+      "[0.010836209170520306,],\n",
+      "[-0.0016270435880869627,],\n",
+      "[-0.47825783491134644,],\n",
+      "[-0.030628995969891548,],\n",
+      "[-0.7577068209648132,],\n",
+      "[-0.2691391110420227,],\n",
+      "[-0.0713268518447876,],\n",
+      "[0.006319376640021801,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22363197803497314,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.09797754883766174,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450813412666321,],\n",
+      "[-0.12847687304019928,],\n",
+      "[0.5475986003875732,],\n",
+      "[-0.9998651146888733,],\n",
+      "[-0.010404622182250023,],\n",
+      "[0.4548550248146057,],\n",
+      "[-0.3821728229522705,],\n",
+      "[0.9758396744728088,],\n",
+      "[0.8241040110588074,],\n",
+      "[-0.9389644265174866,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2237299531698227,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.055045947432518005,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450813412666321,],\n",
+      "[-0.12847687304019928,],\n",
+      "[0.5475986003875732,],\n",
+      "[-0.9998651146888733,],\n",
+      "[-0.010404622182250023,],\n",
+      "[0.4548550248146057,],\n",
+      "[-0.3821728229522705,],\n",
+      "[0.9758396744728088,],\n",
+      "[0.8241040110588074,],\n",
+      "[-0.9389644265174866,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.050201646983623505,],\n",
+      "[0.015817956998944283,],\n",
+      "[0.04328053072094917,],\n",
+      "[-0.006587478332221508,],\n",
+      "[-0.5178449153900146,],\n",
+      "[-0.033460937440395355,],\n",
+      "[-0.8256157040596008,],\n",
+      "[-0.27432331442832947,],\n",
+      "[-0.07434116303920746,],\n",
+      "[0.006354811135679483,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2237299531698227,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4692375361919403,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.645131528377533,],\n",
+      "[-0.12849269807338715,],\n",
+      "[0.5475553274154663,],\n",
+      "[-0.9998584985733032,],\n",
+      "[-0.00988677702844143,],\n",
+      "[0.4548884928226471,],\n",
+      "[-0.3813472092151642,],\n",
+      "[0.9761139750480652,],\n",
+      "[0.8241783380508423,],\n",
+      "[-0.9389708042144775,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22419919073581696,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.010656343773007393,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.645131528377533,],\n",
+      "[-0.12849269807338715,],\n",
+      "[0.5475553274154663,],\n",
+      "[-0.9998584985733032,],\n",
+      "[-0.00988677702844143,],\n",
+      "[0.4548884928226471,],\n",
+      "[-0.3813472092151642,],\n",
+      "[0.9761139750480652,],\n",
+      "[0.8241783380508423,],\n",
+      "[-0.9389708042144775,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.017488019540905952,],\n",
+      "[0.005534573923796415,],\n",
+      "[0.015754882246255875,],\n",
+      "[-0.0024381540715694427,],\n",
+      "[-0.4893954396247864,],\n",
+      "[-0.03174061328172684,],\n",
+      "[-0.7779199481010437,],\n",
+      "[-0.27039048075675964,],\n",
+      "[-0.07175596803426743,],\n",
+      "[0.006308200769126415,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22419919073581696,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2064589112997055,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6451489925384521,],\n",
+      "[-0.12849822640419006,],\n",
+      "[0.5475395917892456,],\n",
+      "[-0.9998560547828674,],\n",
+      "[-0.009397381916642189,],\n",
+      "[0.45492023229599,],\n",
+      "[-0.3805692791938782,],\n",
+      "[0.9763843417167664,],\n",
+      "[0.8242501020431519,],\n",
+      "[-0.9389771223068237,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22440564632415771,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04994392767548561,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6451489925384521,],\n",
+      "[-0.12849822640419006,],\n",
+      "[0.5475395917892456,],\n",
+      "[-0.9998560547828674,],\n",
+      "[-0.009397381916642189,],\n",
+      "[0.45492023229599,],\n",
+      "[-0.3805692791938782,],\n",
+      "[0.9763843417167664,],\n",
+      "[0.8242501020431519,],\n",
+      "[-0.9389771223068237,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.02943040058016777,],\n",
+      "[-0.009419398382306099,],\n",
+      "[-0.028030773624777794,],\n",
+      "[0.004420470912009478,],\n",
+      "[-0.42211809754371643,],\n",
+      "[-0.02690926566720009,],\n",
+      "[-0.6581597924232483,],\n",
+      "[-0.2638132572174072,],\n",
+      "[-0.0680663138628006,],\n",
+      "[0.006264133378863335,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22440564632415771,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4469631314277649,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6451195478439331,],\n",
+      "[-0.12848880887031555,],\n",
+      "[0.54756760597229,],\n",
+      "[-0.9998604655265808,],\n",
+      "[-0.008975263684988022,],\n",
+      "[0.4549471437931061,],\n",
+      "[-0.3799111247062683,],\n",
+      "[0.9766481518745422,],\n",
+      "[0.8243181705474854,],\n",
+      "[-0.9389833807945251,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22395868599414825,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2629704177379608,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6451195478439331,],\n",
+      "[-0.12848880887031555,],\n",
+      "[0.54756760597229,],\n",
+      "[-0.9998604655265808,],\n",
+      "[-0.008975263684988022,],\n",
+      "[0.4549471437931061,],\n",
+      "[-0.3799111247062683,],\n",
+      "[0.9766481518745422,],\n",
+      "[0.8243181705474854,],\n",
+      "[-0.9389833807945251,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.05172353237867355,],\n",
+      "[-0.01685621589422226,],\n",
+      "[-0.052563559263944626,],\n",
+      "[0.008461845107376575,],\n",
+      "[-0.36564144492149353,],\n",
+      "[-0.022409798577427864,],\n",
+      "[-0.5514906644821167,],\n",
+      "[-0.25969648361206055,],\n",
+      "[-0.06633701920509338,],\n",
+      "[0.006253677885979414,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22395868599414825,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.025612473487854,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450678110122681,],\n",
+      "[-0.1284719556570053,],\n",
+      "[0.5476201772689819,],\n",
+      "[-0.9998689293861389,],\n",
+      "[-0.008609621785581112,],\n",
+      "[0.4549695551395416,],\n",
+      "[-0.379359632730484,],\n",
+      "[0.9769078493118286,],\n",
+      "[0.8243845105171204,],\n",
+      "[-0.9389896392822266,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2229330688714981,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.31497204303741455,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450678110122681,],\n",
+      "[-0.1284719556570053,],\n",
+      "[0.5476201772689819,],\n",
+      "[-0.9998689293861389,],\n",
+      "[-0.008609621785581112,],\n",
+      "[0.4549695551395416,],\n",
+      "[-0.379359632730484,],\n",
+      "[0.9769078493118286,],\n",
+      "[0.8243845105171204,],\n",
+      "[-0.9389896392822266,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04286247864365578,],\n",
+      "[-0.014300292357802391,],\n",
+      "[-0.04679263383150101,],\n",
+      "[0.007700815796852112,],\n",
+      "[-0.359080046415329,],\n",
+      "[-0.021388588473200798,],\n",
+      "[-0.5330960750579834,],\n",
+      "[-0.2596643269062042,],\n",
+      "[-0.06701702624559402,],\n",
+      "[0.006269596982747316,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2229330688714981,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.122449517250061,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450249552726746,],\n",
+      "[-0.12845765054225922,],\n",
+      "[0.5476669669151306,],\n",
+      "[-0.9998766183853149,],\n",
+      "[-0.008250541985034943,],\n",
+      "[0.454990953207016,],\n",
+      "[-0.3788265287876129,],\n",
+      "[0.9771674871444702,],\n",
+      "[0.8244515061378479,],\n",
+      "[-0.938995897769928,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22181062400341034,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10043465346097946,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450249552726746,],\n",
+      "[-0.12845765054225922,],\n",
+      "[0.5476669669151306,],\n",
+      "[-0.9998766183853149,],\n",
+      "[-0.008250541985034943,],\n",
+      "[0.454990953207016,],\n",
+      "[-0.3788265287876129,],\n",
+      "[0.9771674871444702,],\n",
+      "[0.8244515061378479,],\n",
+      "[-0.938995897769928,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.018175674602389336,],\n",
+      "[-0.0062254611402750015,],\n",
+      "[-0.021412581205368042,],\n",
+      "[0.0036069010384380817,],\n",
+      "[-0.407909095287323,],\n",
+      "[-0.02457101084291935,],\n",
+      "[-0.6223997473716736,],\n",
+      "[-0.26242536306381226,],\n",
+      "[-0.06887426227331161,],\n",
+      "[0.006292433012276888,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22181062400341034,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6338292360305786,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450067758560181,],\n",
+      "[-0.1284514218568802,],\n",
+      "[0.547688364982605,],\n",
+      "[-0.9998802542686462,],\n",
+      "[-0.007842632941901684,],\n",
+      "[0.45501551032066345,],\n",
+      "[-0.3782041370868683,],\n",
+      "[0.9774299263954163,],\n",
+      "[0.8245204091072083,],\n",
+      "[-0.9390022158622742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22117678821086884,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.008046328090131283,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450067758560181,],\n",
+      "[-0.1284514218568802,],\n",
+      "[0.547688364982605,],\n",
+      "[-0.9998802542686462,],\n",
+      "[-0.007842632941901684,],\n",
+      "[0.45501551032066345,],\n",
+      "[-0.3782041370868683,],\n",
+      "[0.9774299263954163,],\n",
+      "[0.8245204091072083,],\n",
+      "[-0.9390022158622742,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.003841269062831998,],\n",
+      "[0.0013508989941328764,],\n",
+      "[0.004897048696875572,],\n",
+      "[-0.0008451794274151325,],\n",
+      "[-0.4830298125743866,],\n",
+      "[-0.03008333407342434,],\n",
+      "[-0.7717349529266357,],\n",
+      "[-0.26571086049079895,],\n",
+      "[-0.07050079107284546,],\n",
+      "[0.006308000534772873,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22117678821086884,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.17940254509449005,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450105905532837,],\n",
+      "[-0.12845277786254883,],\n",
+      "[0.5476834774017334,],\n",
+      "[-0.9998794198036194,],\n",
+      "[-0.007359602954238653,],\n",
+      "[0.45504558086395264,],\n",
+      "[-0.3774324059486389,],\n",
+      "[0.977695643901825,],\n",
+      "[0.8245909214019775,],\n",
+      "[-0.9390085339546204,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22135619819164276,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.18141812086105347,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450105905532837,],\n",
+      "[-0.12845277786254883,],\n",
+      "[0.5476834774017334,],\n",
+      "[-0.9998794198036194,],\n",
+      "[-0.007359602954238653,],\n",
+      "[0.45504558086395264,],\n",
+      "[-0.3774324059486389,],\n",
+      "[0.977695643901825,],\n",
+      "[0.8245909214019775,],\n",
+      "[-0.9390085339546204,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.013565079309046268,],\n",
+      "[0.0048912144266068935,],\n",
+      "[0.01874416507780552,],\n",
+      "[-0.003317370777949691,],\n",
+      "[-0.5409676432609558,],\n",
+      "[-0.03483515605330467,],\n",
+      "[-0.895791232585907,],\n",
+      "[-0.26792001724243164,],\n",
+      "[-0.071205273270607,],\n",
+      "[0.006312720477581024,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22135619819164276,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.85186368227005,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450241804122925,],\n",
+      "[-0.1284576654434204,],\n",
+      "[0.5476647615432739,],\n",
+      "[-0.999876081943512,],\n",
+      "[-0.006818635389208794,],\n",
+      "[0.45508041977882385,],\n",
+      "[-0.37653660774230957,],\n",
+      "[0.9779635667800903,],\n",
+      "[0.8246621489524841,],\n",
+      "[-0.9390148520469666,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22220806777477264,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.24879279732704163,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450241804122925,],\n",
+      "[-0.1284576654434204,],\n",
+      "[0.5476647615432739,],\n",
+      "[-0.999876081943512,],\n",
+      "[-0.006818635389208794,],\n",
+      "[0.45508041977882385,],\n",
+      "[-0.37653660774230957,],\n",
+      "[0.9779635667800903,],\n",
+      "[0.8246621489524841,],\n",
+      "[-0.9390148520469666,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.01177878025919199,],\n",
+      "[0.00434892438352108,],\n",
+      "[0.017660116776823997,],\n",
+      "[-0.003207189030945301,],\n",
+      "[-0.550780713558197,],\n",
+      "[-0.036200977861881256,],\n",
+      "[-0.9233412742614746,],\n",
+      "[-0.26859745383262634,],\n",
+      "[-0.0710749700665474,],\n",
+      "[0.0063101183623075485,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22220806777477264,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9975838661193848,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.645035982131958,],\n",
+      "[-0.128462016582489,],\n",
+      "[0.5476471185684204,],\n",
+      "[-0.9998728632926941,],\n",
+      "[-0.006267854478210211,],\n",
+      "[0.4551166296005249,],\n",
+      "[-0.375613272190094,],\n",
+      "[0.9782321453094482,],\n",
+      "[0.8247331976890564,],\n",
+      "[-0.9390211701393127,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22320565581321716,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07442580163478851,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.645035982131958,],\n",
+      "[-0.128462016582489,],\n",
+      "[0.5476471185684204,],\n",
+      "[-0.9998728632926941,],\n",
+      "[-0.006267854478210211,],\n",
+      "[0.4551166296005249,],\n",
+      "[-0.375613272190094,],\n",
+      "[0.9782321453094482,],\n",
+      "[0.8247331976890564,],\n",
+      "[-0.9390211701393127,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.004764879588037729,],\n",
+      "[0.0018017018446698785,],\n",
+      "[0.007758892606943846,],\n",
+      "[-0.0014464914565905929,],\n",
+      "[-0.5114453434944153,],\n",
+      "[-0.0334433875977993,],\n",
+      "[-0.8405377268791199,],\n",
+      "[-0.2681308388710022,],\n",
+      "[-0.07058116048574448,],\n",
+      "[0.006305508781224489,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22320565581321716,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5456215739250183,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.64504075050354,],\n",
+      "[-0.12846381962299347,],\n",
+      "[0.5476393699645996,],\n",
+      "[-0.9998714327812195,],\n",
+      "[-0.005756408907473087,],\n",
+      "[0.4551500678062439,],\n",
+      "[-0.37477272748947144,],\n",
+      "[0.978500247001648,],\n",
+      "[0.8248037695884705,],\n",
+      "[-0.9390274882316589,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2237512767314911,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.01163257472217083,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.64504075050354,],\n",
+      "[-0.12846381962299347,],\n",
+      "[0.5476393699645996,],\n",
+      "[-0.9998714327812195,],\n",
+      "[-0.005756408907473087,],\n",
+      "[0.4551500678062439,],\n",
+      "[-0.37477272748947144,],\n",
+      "[0.978500247001648,],\n",
+      "[0.8248037695884705,],\n",
+      "[-0.9390274882316589,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.001390312216244638,],\n",
+      "[-0.0005393580649979413,],\n",
+      "[-0.0024610196705907583,],\n",
+      "[0.00047105943667702377,],\n",
+      "[-0.45073193311691284,],\n",
+      "[-0.02834930084645748,],\n",
+      "[-0.7000488638877869,],\n",
+      "[-0.26717609167099,],\n",
+      "[-0.07015753537416458,],\n",
+      "[0.006302365567535162,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2237512767314911,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.2157091200351715,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450393795967102,],\n",
+      "[-0.1284632831811905,],\n",
+      "[0.5476418137550354,],\n",
+      "[-0.9998719096183777,],\n",
+      "[-0.005305676721036434,],\n",
+      "[0.4551784098148346,],\n",
+      "[-0.3740726709365845,],\n",
+      "[0.9787673950195312,],\n",
+      "[0.8248739242553711,],\n",
+      "[-0.9390338063240051,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22353556752204895,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1652527004480362,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450393795967102,],\n",
+      "[-0.1284632831811905,],\n",
+      "[0.5476418137550354,],\n",
+      "[-0.9998719096183777,],\n",
+      "[-0.005305676721036434,],\n",
+      "[0.4551784098148346,],\n",
+      "[-0.3740726709365845,],\n",
+      "[0.9787673950195312,],\n",
+      "[0.8248739242553711,],\n",
+      "[-0.9390338063240051,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0038609826005995274,],\n",
+      "[-0.0015407073078677058,],\n",
+      "[-0.0074357264675199986,],\n",
+      "[0.0014611316146329045,],\n",
+      "[-0.40700531005859375,],\n",
+      "[-0.02423468977212906,],\n",
+      "[-0.589464545249939,],\n",
+      "[-0.26624050736427307,],\n",
+      "[-0.06999076157808304,],\n",
+      "[0.00630152877420187,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22353556752204895,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8130253553390503,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450355052947998,],\n",
+      "[-0.12846174836158752,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998733997344971,],\n",
+      "[-0.00489867152646184,],\n",
+      "[0.4552026391029358,],\n",
+      "[-0.37348321080207825,],\n",
+      "[0.9790336489677429,],\n",
+      "[0.8249439001083374,],\n",
+      "[-0.9390401244163513,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22272254526615143,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.19129051268100739,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450355052947998,],\n",
+      "[-0.12846174836158752,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998733997344971,],\n",
+      "[-0.00489867152646184,],\n",
+      "[0.4552026391029358,],\n",
+      "[-0.37348321080207825,],\n",
+      "[0.9790336489677429,],\n",
+      "[0.8249439001083374,],\n",
+      "[-0.9390401244163513,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.003057059133425355,],\n",
+      "[-0.001257272670045495,],\n",
+      "[-0.006409129593521357,],\n",
+      "[0.0012927564093843102,],\n",
+      "[-0.40526750683784485,],\n",
+      "[-0.023685291409492493,],\n",
+      "[-0.5777440667152405,],\n",
+      "[-0.2655865252017975,],\n",
+      "[-0.07004515081644058,],\n",
+      "[0.0063021923415362835,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22272254526615143,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8747343420982361,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450324654579163,],\n",
+      "[-0.12846049666404724,],\n",
+      "[0.5476557016372681,],\n",
+      "[-0.9998747110366821,],\n",
+      "[-0.004493404179811478,],\n",
+      "[0.455226331949234,],\n",
+      "[-0.3729054629802704,],\n",
+      "[0.9792992472648621,],\n",
+      "[0.8250139355659485,],\n",
+      "[-0.9390464425086975,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22184781730175018,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0317428857088089,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450324654579163,],\n",
+      "[-0.12846049666404724,],\n",
+      "[0.5476557016372681,],\n",
+      "[-0.9998747110366821,],\n",
+      "[-0.004493404179811478,],\n",
+      "[0.455226331949234,],\n",
+      "[-0.3729054629802704,],\n",
+      "[0.9792992472648621,],\n",
+      "[0.8250139355659485,],\n",
+      "[-0.9390464425086975,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0009157882304862142,],\n",
+      "[-0.0003882735036313534,],\n",
+      "[-0.002090366091579199,],\n",
+      "[0.00043282631668262184,],\n",
+      "[-0.4430755376815796,],\n",
+      "[-0.026886621490120888,],\n",
+      "[-0.6731712222099304,],\n",
+      "[-0.2653411328792572,],\n",
+      "[-0.07018530368804932,],\n",
+      "[0.006303164642304182,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22184781730175018,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3563311994075775,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450315713882446,],\n",
+      "[-0.1284601092338562,],\n",
+      "[0.5476577877998352,],\n",
+      "[-0.9998751282691956,],\n",
+      "[-0.004050328396260738,],\n",
+      "[0.4552532136440277,],\n",
+      "[-0.3722322881221771,],\n",
+      "[0.9795646071434021,],\n",
+      "[0.8250841498374939,],\n",
+      "[-0.9390527606010437,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2214914858341217,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04682521894574165,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450315713882446,],\n",
+      "[-0.1284601092338562,],\n",
+      "[0.5476577877998352,],\n",
+      "[-0.9998751282691956,],\n",
+      "[-0.004050328396260738,],\n",
+      "[0.4552532136440277,],\n",
+      "[-0.3722322881221771,],\n",
+      "[0.9795646071434021,],\n",
+      "[0.8250841498374939,],\n",
+      "[-0.9390527606010437,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0008176066912710667,],\n",
+      "[0.00035683048190549016,],\n",
+      "[0.0020313281565904617,],\n",
+      "[-0.0004319036961533129,],\n",
+      "[-0.4948274791240692,],\n",
+      "[-0.03178078308701515,],\n",
+      "[-0.8187181353569031,],\n",
+      "[-0.2655790150165558,],\n",
+      "[-0.07029621303081512,],\n",
+      "[0.006303820293396711,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2214914858341217,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4327831566333771,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450324058532715,],\n",
+      "[-0.12846046686172485,],\n",
+      "[0.5476557612419128,],\n",
+      "[-0.9998747110366821,],\n",
+      "[-0.003555500879883766,],\n",
+      "[0.455284982919693,],\n",
+      "[-0.37141355872154236,],\n",
+      "[0.9798302054405212,],\n",
+      "[0.8251544237136841,],\n",
+      "[-0.9390590786933899,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22192427515983582,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.25873568654060364,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450324058532715,],\n",
+      "[-0.12846046686172485,],\n",
+      "[0.5476557612419128,],\n",
+      "[-0.9998747110366821,],\n",
+      "[-0.003555500879883766,],\n",
+      "[0.455284982919693,],\n",
+      "[-0.37141355872154236,],\n",
+      "[0.9798302054405212,],\n",
+      "[0.8251544237136841,],\n",
+      "[-0.9390590786933899,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0014123632572591305,],\n",
+      "[0.0006330818869173527,],\n",
+      "[0.0038174442015588284,],\n",
+      "[-0.0008338932530023158,],\n",
+      "[-0.529353141784668,],\n",
+      "[-0.0354853980243206,],\n",
+      "[-0.9267815947532654,],\n",
+      "[-0.26626402139663696,],\n",
+      "[-0.07033346593379974,],\n",
+      "[0.006303918082267046,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22192427515983582,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0173205137252808,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450338363647461,],\n",
+      "[-0.128461092710495,],\n",
+      "[0.5476519465446472,],\n",
+      "[-0.9998738765716553,],\n",
+      "[-0.0030261478386819363,],\n",
+      "[0.45532047748565674,],\n",
+      "[-0.3704867660999298,],\n",
+      "[0.9800964593887329,],\n",
+      "[0.825224757194519,],\n",
+      "[-0.9390653967857361,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.222941592335701,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.27298495173454285,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450338363647461,],\n",
+      "[-0.128461092710495,],\n",
+      "[0.5476519465446472,],\n",
+      "[-0.9998738765716553,],\n",
+      "[-0.0030261478386819363,],\n",
+      "[0.45532047748565674,],\n",
+      "[-0.3704867660999298,],\n",
+      "[0.9800964593887329,],\n",
+      "[0.825224757194519,],\n",
+      "[-0.9390653967857361,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0010658574756234884,],\n",
+      "[0.0004898727056570351,],\n",
+      "[0.0031328133773058653,],\n",
+      "[-0.0007033581496216357,],\n",
+      "[-0.5282726287841797,],\n",
+      "[-0.03589014708995819,],\n",
+      "[-0.9322484135627747,],\n",
+      "[-0.26718392968177795,],\n",
+      "[-0.07031163573265076,],\n",
+      "[0.006303710862994194,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.222941592335701,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.044958472251892,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.645034909248352,],\n",
+      "[-0.1284615844488144,],\n",
+      "[0.5476487874984741,],\n",
+      "[-0.999873161315918,],\n",
+      "[-0.0024978751316666603,],\n",
+      "[0.4553563594818115,],\n",
+      "[-0.3695545196533203,],\n",
+      "[0.980363667011261,],\n",
+      "[0.825295090675354,],\n",
+      "[-0.9390717148780823,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22398655116558075,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.06265927106142044,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.645034909248352,],\n",
+      "[-0.1284615844488144,],\n",
+      "[0.5476487874984741,],\n",
+      "[-0.999873161315918,],\n",
+      "[-0.0024978751316666603,],\n",
+      "[0.4553563594818115,],\n",
+      "[-0.3695545196533203,],\n",
+      "[0.980363667011261,],\n",
+      "[0.825295090675354,],\n",
+      "[-0.9390717148780823,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0003750871983356774,],\n",
+      "[0.0001767490030033514,],\n",
+      "[0.0011987974867224693,],\n",
+      "[-0.00027664072695188224,],\n",
+      "[-0.4952956736087799,],\n",
+      "[-0.032831549644470215,],\n",
+      "[-0.8322745561599731,],\n",
+      "[-0.26795002818107605,],\n",
+      "[-0.0702701061964035,],\n",
+      "[0.0063034771010279655,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22398655116558075,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5006371140480042,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450352668762207,],\n",
+      "[-0.12846176326274872,],\n",
+      "[0.5476475954055786,],\n",
+      "[-0.9998728632926941,],\n",
+      "[-0.0020025793928653,],\n",
+      "[0.45538920164108276,],\n",
+      "[-0.36872223019599915,],\n",
+      "[0.9806315898895264,],\n",
+      "[0.8253653645515442,],\n",
+      "[-0.9390780329704285,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22448718547821045,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.02015414647758007,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450352668762207,],\n",
+      "[-0.12846176326274872,],\n",
+      "[0.5476475954055786,],\n",
+      "[-0.9998728632926941,],\n",
+      "[-0.0020025793928653,],\n",
+      "[0.45538920164108276,],\n",
+      "[-0.36872223019599915,],\n",
+      "[0.9806315898895264,],\n",
+      "[0.8253653645515442,],\n",
+      "[-0.9390780329704285,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00015621916099917144,],\n",
+      "[-7.559332152595744e-05,],\n",
+      "[-0.0005430780001915991,],\n",
+      "[0.00012876994151156396,],\n",
+      "[-0.4524793028831482,],\n",
+      "[-0.028277572244405746,],\n",
+      "[-0.688215970993042,],\n",
+      "[-0.2681996822357178,],\n",
+      "[-0.0702376514673233,],\n",
+      "[0.006303321570158005,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22448718547821045,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.28393077850341797,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450350880622864,],\n",
+      "[-0.12846168875694275,],\n",
+      "[0.5476481318473816,],\n",
+      "[-0.9998729825019836,],\n",
+      "[-0.0015501000452786684,],\n",
+      "[0.4554174840450287,],\n",
+      "[-0.3680340051651001,],\n",
+      "[0.9808998107910156,],\n",
+      "[0.8254355788230896,],\n",
+      "[-0.9390843510627747,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22420325875282288,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.17285604774951935,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450350880622864,],\n",
+      "[-0.12846168875694275,],\n",
+      "[0.5476481318473816,],\n",
+      "[-0.9998729825019836,],\n",
+      "[-0.0015501000452786684,],\n",
+      "[0.4554174840450287,],\n",
+      "[-0.3680340051651001,],\n",
+      "[0.9808998107910156,],\n",
+      "[0.8254355788230896,],\n",
+      "[-0.9390843510627747,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00033591638202778995,],\n",
+      "[-0.00016729462367948145,],\n",
+      "[-0.0012708373833447695,],\n",
+      "[0.0003094485145993531,],\n",
+      "[-0.4258587062358856,],\n",
+      "[-0.025105876848101616,],\n",
+      "[-0.5877054333686829,],\n",
+      "[-0.26781174540519714,],\n",
+      "[-0.07022783160209656,],\n",
+      "[0.006303281057626009,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22420325875282288,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8315191864967346,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450347304344177,],\n",
+      "[-0.12846152484416962,],\n",
+      "[0.5476493835449219,],\n",
+      "[-0.9998732805252075,],\n",
+      "[-0.0011242413893342018,],\n",
+      "[0.4554425776004791,],\n",
+      "[-0.36744630336761475,],\n",
+      "[0.9811676144599915,],\n",
+      "[0.825505793094635,],\n",
+      "[-0.9390906691551208,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2233717441558838,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1636974960565567,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450347304344177,],\n",
+      "[-0.12846152484416962,],\n",
+      "[0.5476493835449219,],\n",
+      "[-0.9998732805252075,],\n",
+      "[-0.0011242413893342018,],\n",
+      "[0.4554425776004791,],\n",
+      "[-0.36744630336761475,],\n",
+      "[0.9811676144599915,],\n",
+      "[0.825505793094635,],\n",
+      "[-0.9390906691551208,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00023999306722544134,],\n",
+      "[-0.0001232052018167451,],\n",
+      "[-0.0009884241735562682,],\n",
+      "[0.0002470649778842926,],\n",
+      "[-0.4292173981666565,],\n",
+      "[-0.025177819654345512,],\n",
+      "[-0.5914765000343323,],\n",
+      "[-0.26702263951301575,],\n",
+      "[-0.07023276388645172,],\n",
+      "[0.00630333973094821,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2233717441558838,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8091921806335449,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450344920158386,],\n",
+      "[-0.12846140563488007,],\n",
+      "[0.5476503968238831,],\n",
+      "[-0.9998735189437866,],\n",
+      "[-0.0006950239767320454,],\n",
+      "[0.45546776056289673,],\n",
+      "[-0.36685481667518616,],\n",
+      "[0.9814346432685852,],\n",
+      "[0.8255760073661804,],\n",
+      "[-0.939096987247467,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22256255149841309,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.013294732198119164,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450344920158386,],\n",
+      "[-0.12846140563488007,],\n",
+      "[0.5476503968238831,],\n",
+      "[-0.9998735189437866,],\n",
+      "[-0.0006950239767320454,],\n",
+      "[0.45546776056289673,],\n",
+      "[-0.36685481667518616,],\n",
+      "[0.9814346432685852,],\n",
+      "[0.8255760073661804,],\n",
+      "[-0.939096987247467,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[5.020894604967907e-05,],\n",
+      "[-2.656792821653653e-05,],\n",
+      "[-0.0002251067926408723,],\n",
+      "[5.775982208433561e-05,],\n",
+      "[-0.4570411443710327,],\n",
+      "[-0.028228024020791054,],\n",
+      "[-0.6970937848091125,],\n",
+      "[-0.2663212716579437,],\n",
+      "[-0.07024374604225159,],\n",
+      "[0.006303394678980112,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22256255149841309,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.23060579597949982,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450344324111938,],\n",
+      "[-0.12846137583255768,],\n",
+      "[0.5476506352424622,],\n",
+      "[-0.9998735785484314,],\n",
+      "[-0.00023798280744813383,],\n",
+      "[0.4554959833621979,],\n",
+      "[-0.3661577105522156,],\n",
+      "[0.9817009568214417,],\n",
+      "[0.8256462216377258,],\n",
+      "[-0.9391033053398132,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2223319411277771,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07620859891176224,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450344324111938,],\n",
+      "[-0.12846137583255768,],\n",
+      "[0.5476506352424622,],\n",
+      "[-0.9998735785484314,],\n",
+      "[-0.00023798280744813383,],\n",
+      "[0.4554959833621979,],\n",
+      "[-0.3661577105522156,],\n",
+      "[0.9817009568214417,],\n",
+      "[0.8256462216377258,],\n",
+      "[-0.9391033053398132,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.824664109852165e-05,],\n",
+      "[4.8050813347799703e-05,],\n",
+      "[0.00043049216037616134,],\n",
+      "[-0.00011343746155034751,],\n",
+      "[-0.4903267025947571,],\n",
+      "[-0.032266728579998016,],\n",
+      "[-0.8401437401771545,],\n",
+      "[-0.26618316769599915,],\n",
+      "[-0.07025190442800522,],\n",
+      "[0.006303396075963974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2223319411277771,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5521180033683777,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450344920158386,],\n",
+      "[-0.12846142053604126,],\n",
+      "[0.5476502180099487,],\n",
+      "[-0.9998734593391418,],\n",
+      "[0.00025234391796402633,],\n",
+      "[0.45552825927734375,],\n",
+      "[-0.36531755328178406,],\n",
+      "[0.9819671511650085,],\n",
+      "[0.825716495513916,],\n",
+      "[-0.9391096234321594,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22288405895233154,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2820696234703064,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450344920158386,],\n",
+      "[-0.12846142053604126,],\n",
+      "[0.5476502180099487,],\n",
+      "[-0.9998734593391418,],\n",
+      "[0.00025234391796402633,],\n",
+      "[0.45552825927734375,],\n",
+      "[-0.36531755328178406,],\n",
+      "[0.9819671511650085,],\n",
+      "[0.825716495513916,],\n",
+      "[-0.9391096234321594,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0001246299216290936,],\n",
+      "[6.967649824218825e-05,],\n",
+      "[0.0006611105054616928,],\n",
+      "[-0.00017901173850987107,],\n",
+      "[-0.5092874765396118,],\n",
+      "[-0.03489263728260994,],\n",
+      "[-0.9330321550369263,],\n",
+      "[-0.26676782965660095,],\n",
+      "[-0.07025346159934998,],\n",
+      "[0.00630339328199625,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22288405895233154,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.06220281124115,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846149504184723,],\n",
+      "[0.5476495623588562,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.0007616314105689526,],\n",
+      "[0.45556315779685974,],\n",
+      "[-0.3643845319747925,],\n",
+      "[0.9822339415550232,],\n",
+      "[0.8257867693901062,],\n",
+      "[-0.9391159415245056,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22394625842571259,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.24402298033237457,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846149504184723,],\n",
+      "[0.5476495623588562,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.0007616314105689526,],\n",
+      "[0.45556315779685974,],\n",
+      "[-0.3643845319747925,],\n",
+      "[0.9822339415550232,],\n",
+      "[0.8257867693901062,],\n",
+      "[-0.9391159415245056,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.509431063430384e-05,],\n",
+      "[4.877440369455144e-05,],\n",
+      "[0.000490623468067497,],\n",
+      "[-0.00013656697410624474,],\n",
+      "[-0.5044013261795044,],\n",
+      "[-0.03463433310389519,],\n",
+      "[-0.9190240502357483,],\n",
+      "[-0.2678058445453644,],\n",
+      "[-0.07025132328271866,],\n",
+      "[0.006303390022367239,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22394625842571259,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9879752993583679,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.645034670829773,],\n",
+      "[-0.1284615397453308,],\n",
+      "[0.547649085521698,],\n",
+      "[-0.999873161315918,],\n",
+      "[0.0012660326901823282,],\n",
+      "[0.45559778809547424,],\n",
+      "[-0.3634655177593231,],\n",
+      "[0.982501745223999,],\n",
+      "[0.8258570432662964,],\n",
+      "[-0.9391222596168518,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22493423521518707,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03467393293976784,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.645034670829773,],\n",
+      "[-0.1284615397453308,],\n",
+      "[0.547649085521698,],\n",
+      "[-0.999873161315918,],\n",
+      "[0.0012660326901823282,],\n",
+      "[0.45559778809547424,],\n",
+      "[-0.3634655177593231,],\n",
+      "[0.982501745223999,],\n",
+      "[0.8258570432662964,],\n",
+      "[-0.9391222596168518,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.354596472287085e-05,],\n",
+      "[1.3839605344401207e-05,],\n",
+      "[0.0001475607423344627,],\n",
+      "[-4.2221654439345e-05,],\n",
+      "[-0.4806879758834839,],\n",
+      "[-0.03176113963127136,],\n",
+      "[-0.8076106905937195,],\n",
+      "[-0.26874586939811707,],\n",
+      "[-0.07024841010570526,],\n",
+      "[0.0063033876940608025,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22493423521518707,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3724187910556793,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.645034670829773,],\n",
+      "[-0.128461554646492,],\n",
+      "[0.5476489663124084,],\n",
+      "[-0.9998731017112732,],\n",
+      "[0.0017467206344008446,],\n",
+      "[0.4556295573711395,],\n",
+      "[-0.36265790462493896,],\n",
+      "[0.9827705025672913,],\n",
+      "[0.8259273171424866,],\n",
+      "[-0.939128577709198,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2253066599369049,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04259604588150978,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.645034670829773,],\n",
+      "[-0.128461554646492,],\n",
+      "[0.5476489663124084,],\n",
+      "[-0.9998731017112732,],\n",
+      "[0.0017467206344008446,],\n",
+      "[0.4556295573711395,],\n",
+      "[-0.36265790462493896,],\n",
+      "[0.9827705025672913,],\n",
+      "[0.8259273171424866,],\n",
+      "[-0.939128577709198,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.915667235152796e-05,],\n",
+      "[-1.15664915938396e-05,],\n",
+      "[-0.00013055156159680337,],\n",
+      "[3.8379210309358314e-05,],\n",
+      "[-0.4537178575992584,],\n",
+      "[-0.028123702853918076,],\n",
+      "[-0.6671080589294434,],\n",
+      "[-0.2690794765949249,],\n",
+      "[-0.07024618983268738,],\n",
+      "[0.006303387228399515,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2253066599369049,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.4127763509750366,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.645034670829773,],\n",
+      "[-0.1284615397453308,],\n",
+      "[0.547649085521698,],\n",
+      "[-0.999873161315918,],\n",
+      "[0.002200438641011715,],\n",
+      "[0.4556576907634735,],\n",
+      "[-0.3619908094406128,],\n",
+      "[0.9830395579338074,],\n",
+      "[0.8259975910186768,],\n",
+      "[-0.9391348958015442,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22489388287067413,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.199314683675766,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.645034670829773,],\n",
+      "[-0.1284615397453308,],\n",
+      "[0.547649085521698,],\n",
+      "[-0.999873161315918,],\n",
+      "[0.002200438641011715,],\n",
+      "[0.4556576907634735,],\n",
+      "[-0.3619908094406128,],\n",
+      "[0.9830395579338074,],\n",
+      "[0.8259975910186768,],\n",
+      "[-0.9391348958015442,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.0417004381888546e-05,],\n",
+      "[-1.890791827463545e-05,],\n",
+      "[-0.0002255572035210207,],\n",
+      "[6.808261241531e-05,],\n",
+      "[-0.4394850730895996,],\n",
+      "[-0.025962093845009804,],\n",
+      "[-0.5813661217689514,],\n",
+      "[-0.2686518728733063,],\n",
+      "[-0.07024542987346649,],\n",
+      "[0.0063033876940608025,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22489388287067413,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8928941488265991,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846152484416962,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.0026399237103760242,],\n",
+      "[0.4556836485862732,],\n",
+      "[-0.361409455537796,],\n",
+      "[0.9833081960678101,],\n",
+      "[0.8260678648948669,],\n",
+      "[-0.9391412138938904,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2240009903907776,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15149396657943726,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846152484416962,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.0026399237103760242,],\n",
+      "[0.4556836485862732,],\n",
+      "[-0.361409455537796,],\n",
+      "[0.9833081960678101,],\n",
+      "[0.8260678648948669,],\n",
+      "[-0.9391412138938904,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.9464947399683297e-05,],\n",
+      "[-1.2474286450014915e-05,],\n",
+      "[-0.0001571208704262972,],\n",
+      "[4.867495226790197e-05,],\n",
+      "[-0.4446001648902893,],\n",
+      "[-0.02645466849207878,],\n",
+      "[-0.601081907749176,],\n",
+      "[-0.26777100563049316,],\n",
+      "[-0.0702461525797844,],\n",
+      "[0.00630338815972209,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2240009903907776,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7784440517425537,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476495027542114,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.0030845238361507654,],\n",
+      "[0.45571011304855347,],\n",
+      "[-0.3608083724975586,],\n",
+      "[0.9835759401321411,],\n",
+      "[0.8261381387710571,],\n",
+      "[-0.9391475319862366,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.223222553730011,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.004990397021174431,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476495027542114,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.0030845238361507654,],\n",
+      "[0.45571011304855347,],\n",
+      "[-0.3608083724975586,],\n",
+      "[0.9835759401321411,],\n",
+      "[0.8261381387710571,],\n",
+      "[-0.9391475319862366,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.593159024399938e-06,],\n",
+      "[-1.7127060800703475e-06,],\n",
+      "[-2.2782829546486028e-05,],\n",
+      "[7.2444777288183104e-06,],\n",
+      "[-0.4634961485862732,],\n",
+      "[-0.02908431366086006,],\n",
+      "[-0.7143874764442444,],\n",
+      "[-0.26701053977012634,],\n",
+      "[-0.07024688273668289,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.223222553730011,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.14128537476062775,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476495027542114,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.0035480200313031673,],\n",
+      "[0.45573920011520386,],\n",
+      "[-0.36009398102760315,],\n",
+      "[0.9838429689407349,],\n",
+      "[0.8262084126472473,],\n",
+      "[-0.9391538500785828,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22308126091957092,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09880063682794571,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476495027542114,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.0035480200313031673,],\n",
+      "[0.45573920011520386,],\n",
+      "[-0.36009398102760315,],\n",
+      "[0.9838429689407349,],\n",
+      "[0.8262084126472473,],\n",
+      "[-0.9391538500785828,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.469306521874387e-06,],\n",
+      "[5.754508947575232e-06,],\n",
+      "[8.094507938949391e-05,],\n",
+      "[-2.6433021048433147e-05,],\n",
+      "[-0.4835025668144226,],\n",
+      "[-0.03213993087410927,],\n",
+      "[-0.8516151905059814,],\n",
+      "[-0.2668841779232025,],\n",
+      "[-0.07024762779474258,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22308126091957092,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6286506652832031,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476494431495667,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.004031522665172815,],\n",
+      "[0.4557713270187378,],\n",
+      "[-0.35924237966537476,],\n",
+      "[0.9841098785400391,],\n",
+      "[0.8262786865234375,],\n",
+      "[-0.939160168170929,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22370991110801697,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2810807526111603,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476494431495667,],\n",
+      "[-0.9998732805252075,],\n",
+      "[0.004031522665172815,],\n",
+      "[0.4557713270187378,],\n",
+      "[-0.35924237966537476,],\n",
+      "[0.9841098785400391,],\n",
+      "[0.8262786865234375,],\n",
+      "[-0.939160168170929,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.0485506209079176e-05,],\n",
+      "[7.313202331715729e-06,],\n",
+      "[0.00010894675506278872,],\n",
+      "[-3.656011540442705e-05,],\n",
+      "[-0.49301043152809143,],\n",
+      "[-0.03381117060780525,],\n",
+      "[-0.9272518754005432,],\n",
+      "[-0.26751577854156494,],\n",
+      "[-0.07024763524532318,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22370991110801697,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0603424310684204,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.004524533171206713,],\n",
+      "[0.4558051526546478,],\n",
+      "[-0.3583151400089264,],\n",
+      "[0.984377384185791,],\n",
+      "[0.8263489603996277,],\n",
+      "[-0.9391664862632751,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22477024793624878,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1972280740737915,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.004524533171206713,],\n",
+      "[0.4558051526546478,],\n",
+      "[-0.3583151400089264,],\n",
+      "[0.984377384185791,],\n",
+      "[0.8263489603996277,],\n",
+      "[-0.9391664862632751,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.4470905272173695e-06,],\n",
+      "[4.610046289599268e-06,],\n",
+      "[7.279638521140441e-05,],\n",
+      "[-2.5112925868597813e-05,],\n",
+      "[-0.4876771569252014,],\n",
+      "[-0.03320130705833435,],\n",
+      "[-0.8952072858810425,],\n",
+      "[-0.26857009530067444,],\n",
+      "[-0.07024689763784409,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22477024793624878,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.8882062435150146,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.005012210458517075,],\n",
+      "[0.4558383524417877,],\n",
+      "[-0.3574199378490448,],\n",
+      "[0.9846459627151489,],\n",
+      "[0.8264192342758179,],\n",
+      "[-0.9391728043556213,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22565846145153046,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.011391717940568924,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.005012210458517075,],\n",
+      "[0.4558383524417877,],\n",
+      "[-0.3574199378490448,],\n",
+      "[0.9846459627151489,],\n",
+      "[0.8264192342758179,],\n",
+      "[-0.9391728043556213,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.1373114148227614e-06,],\n",
+      "[8.341658599420043e-07,],\n",
+      "[1.395737945131259e-05,],\n",
+      "[-4.949060439685127e-06,],\n",
+      "[-0.4720735549926758,],\n",
+      "[-0.030799098312854767,],\n",
+      "[-0.7777188420295715,],\n",
+      "[-0.269452303647995,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22565846145153046,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2134639322757721,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.005484283901751041,],\n",
+      "[0.4558691382408142,],\n",
+      "[-0.3566422164440155,],\n",
+      "[0.9849154353141785,],\n",
+      "[0.8264895081520081,],\n",
+      "[-0.9391791224479675,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22587192058563232,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0777272954583168,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.005484283901751041,],\n",
+      "[0.4558691382408142,],\n",
+      "[-0.3566422164440155,],\n",
+      "[0.9849154353141785,],\n",
+      "[0.8264895081520081,],\n",
+      "[-0.9391791224479675,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.1805915366712725e-06,],\n",
+      "[-1.6436641772088478e-06,],\n",
+      "[-2.9102049666107632e-05,],\n",
+      "[1.0600439964036923e-05,],\n",
+      "[-0.45630350708961487,],\n",
+      "[-0.028142914175987244,],\n",
+      "[-0.6465576887130737,],\n",
+      "[-0.26965978741645813,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22587192058563232,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5575921535491943,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.005940587259829044,],\n",
+      "[0.4558972716331482,],\n",
+      "[-0.35599565505981445,],\n",
+      "[0.9851850867271423,],\n",
+      "[0.8265597820281982,],\n",
+      "[-0.9391854405403137,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22531433403491974,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2303956151008606,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476492643356323,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.005940587259829044,],\n",
+      "[0.4558972716331482,],\n",
+      "[-0.35599565505981445,],\n",
+      "[0.9851850867271423,],\n",
+      "[0.8265597820281982,],\n",
+      "[-0.9391854405403137,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.75567026619683e-06,],\n",
+      "[-2.1393716451711953e-06,],\n",
+      "[-4.002067362307571e-05,],\n",
+      "[1.4965075934014749e-05,],\n",
+      "[-0.4494834840297699,],\n",
+      "[-0.02682598866522312,],\n",
+      "[-0.5777423977851868,],\n",
+      "[-0.26909923553466797,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22531433403491974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9599911570549011,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.006390070542693138,],\n",
+      "[0.4559240937232971,],\n",
+      "[-0.3554179072380066,],\n",
+      "[0.9854542016983032,],\n",
+      "[0.8266300559043884,],\n",
+      "[-0.9391917586326599,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2243543416261673,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14142511785030365,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.006390070542693138,],\n",
+      "[0.4559240937232971,],\n",
+      "[-0.3554179072380066,],\n",
+      "[0.9854542016983032,],\n",
+      "[0.8266300559043884,],\n",
+      "[-0.9391917586326599,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.5847290342207998e-06,],\n",
+      "[-1.2686190302702016e-06,],\n",
+      "[-2.505348857084755e-05,],\n",
+      "[9.61402474786155e-06,],\n",
+      "[-0.4543159604072571,],\n",
+      "[-0.027503645047545433,],\n",
+      "[-0.6115939617156982,],\n",
+      "[-0.2681422531604767,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2243543416261673,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7521312236785889,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.006844386458396912,],\n",
+      "[0.45595160126686096,],\n",
+      "[-0.3548063039779663,],\n",
+      "[0.9857223629951477,],\n",
+      "[0.8267003297805786,],\n",
+      "[-0.9391980767250061,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22360220551490784,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0010009314864873886,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.006844386458396912,],\n",
+      "[0.45595160126686096,],\n",
+      "[-0.3548063039779663,],\n",
+      "[0.9857223629951477,],\n",
+      "[0.8267003297805786,],\n",
+      "[-0.9391980767250061,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[9.785807009166092e-08,],\n",
+      "[-8.072878188158938e-08,],\n",
+      "[-1.6837842622408061e-06,],\n",
+      "[6.631982500948652e-07,],\n",
+      "[-0.46638649702072144,],\n",
+      "[-0.029589438810944557,],\n",
+      "[-0.7289291024208069,],\n",
+      "[-0.2673931121826172,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22360220551490784,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.06327489018440247,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.007310772780328989,],\n",
+      "[0.45598119497299194,],\n",
+      "[-0.35407736897468567,],\n",
+      "[0.9859897494316101,],\n",
+      "[0.8267706036567688,],\n",
+      "[-0.9392043948173523,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22353893518447876,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11707376688718796,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.007310772780328989,],\n",
+      "[0.45598119497299194,],\n",
+      "[-0.35407736897468567,],\n",
+      "[0.9859897494316101,],\n",
+      "[0.8267706036567688,],\n",
+      "[-0.9392043948173523,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.768318255330087e-07,],\n",
+      "[6.591222359020321e-07,],\n",
+      "[1.4539114090439398e-05,],\n",
+      "[-5.881258857698413e-06,],\n",
+      "[-0.4778074324131012,],\n",
+      "[-0.031739551573991776,],\n",
+      "[-0.8568423390388489,],\n",
+      "[-0.2673328220844269,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22353893518447876,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.684320867061615,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.007788580376654863,],\n",
+      "[0.45601293444633484,],\n",
+      "[-0.35322052240371704,],\n",
+      "[0.9862570762634277,],\n",
+      "[0.826840877532959,],\n",
+      "[-0.9392107129096985,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22422325611114502,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2665766179561615,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.007788580376654863,],\n",
+      "[0.45601293444633484,],\n",
+      "[-0.35322052240371704,],\n",
+      "[0.9862570762634277,],\n",
+      "[0.826840877532959,],\n",
+      "[-0.9392107129096985,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.604176287008158e-07,],\n",
+      "[7.492649274354335e-07,],\n",
+      "[1.7505104551673867e-05,],\n",
+      "[-7.276961241586832e-06,],\n",
+      "[-0.48213985562324524,],\n",
+      "[-0.03269589692354202,],\n",
+      "[-0.9138787388801575,],\n",
+      "[-0.26801714301109314,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22422325611114502,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0326204299926758,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.008270720019936562,],\n",
+      "[0.45604562759399414,],\n",
+      "[-0.35230663418769836,],\n",
+      "[0.9865251183509827,],\n",
+      "[0.8269111514091492,],\n",
+      "[-0.9392170310020447,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22525587677955627,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14698240160942078,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.008270720019936562,],\n",
+      "[0.45604562759399414,],\n",
+      "[-0.35230663418769836,],\n",
+      "[0.9865251183509827,],\n",
+      "[0.8269111514091492,],\n",
+      "[-0.9392170310020447,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-4.6895766558918694e-07,],\n",
+      "[4.187108402220474e-07,],\n",
+      "[1.0368146831751801e-05,],\n",
+      "[-4.430679382494418e-06,],\n",
+      "[-0.4777168929576874,],\n",
+      "[-0.03198555111885071,],\n",
+      "[-0.8665095567703247,],\n",
+      "[-0.2690467834472656,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22525587677955627,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7667660117149353,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.008748437277972698,],\n",
+      "[0.45607760548591614,],\n",
+      "[-0.3514401316642761,],\n",
+      "[0.9867941737174988,],\n",
+      "[0.8269814252853394,],\n",
+      "[-0.9392233490943909,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2260226458311081,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0005024892743676901,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.008748437277972698,],\n",
+      "[0.45607760548591614,],\n",
+      "[-0.3514401316642761,],\n",
+      "[0.9867941737174988,],\n",
+      "[0.8269814252853394,],\n",
+      "[-0.9392233490943909,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.0126366351291836e-08,],\n",
+      "[1.8438569782119885e-08,],\n",
+      "[4.836635980609572e-07,],\n",
+      "[-2.1242104253360594e-07,],\n",
+      "[-0.4680537283420563,],\n",
+      "[-0.03014202043414116,],\n",
+      "[-0.7485589385032654,],\n",
+      "[-0.26981353759765625,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2260226458311081,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.0448325015604496,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.009216491132974625,],\n",
+      "[0.4561077356338501,],\n",
+      "[-0.3506915867328644,],\n",
+      "[0.987064003944397,],\n",
+      "[0.8270516991615295,],\n",
+      "[-0.9392296671867371,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22606748342514038,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.12207801640033722,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.009216491132974625,],\n",
+      "[0.4561077356338501,],\n",
+      "[-0.3506915867328644,],\n",
+      "[0.987064003944397,],\n",
+      "[0.8270516991615295,],\n",
+      "[-0.9392296671867371,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.3026274220683263e-07,],\n",
+      "[-2.1689987761419616e-07,],\n",
+      "[-6.018398380547296e-06,],\n",
+      "[2.714866468522814e-06,],\n",
+      "[-0.45931392908096313,],\n",
+      "[-0.02833441272377968,],\n",
+      "[-0.6307750940322876,],\n",
+      "[-0.26985836029052734,],\n",
+      "[-0.07024690508842468,],\n",
+      "[0.006303388625383377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22606748342514038,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.698793351650238,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.00967580545693636,],\n",
+      "[0.4561360776424408,],\n",
+      "[-0.3500608205795288,],\n",
+      "[0.9873338341712952,],\n",
+      "[0.8271219730377197,],\n",
+      "[-0.9392359852790833,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2253686934709549,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [6/10], Loss: 0.1221\n",
+      "loss_antes tensor([0.6033365726470947,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6450346112251282,],\n",
+      "[-0.12846150994300842,],\n",
+      "[0.5476493239402771,],\n",
+      "[-0.9998732209205627,],\n",
+      "[0.00967580545693636,],\n",
+      "[0.4561360776424408,],\n",
+      "[-0.3500608205795288,],\n",
+      "[0.9873338341712952,],\n",
+      "[0.8271219730377197,],\n",
+      "[-0.9392359852790833,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.31440263986587524,],\n",
+      "[-0.13579578697681427,],\n",
+      "[-0.4173547625541687,],\n",
+      "[0.06591973453760147,],\n",
+      "[-0.28592395782470703,],\n",
+      "[-0.018915561959147453,],\n",
+      "[-0.4618138074874878,],\n",
+      "[-0.18646888434886932,],\n",
+      "[-0.047146379947662354,],\n",
+      "[0.004324747249484062,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2253686934709549,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.5534965991973877,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.644720196723938,],\n",
+      "[-0.1283257156610489,],\n",
+      "[0.5480666756629944,],\n",
+      "[-0.9999391436576843,],\n",
+      "[0.009961729869246483,],\n",
+      "[0.456155002117157,],\n",
+      "[-0.3495990037918091,],\n",
+      "[0.9875202775001526,],\n",
+      "[0.8271691203117371,],\n",
+      "[-0.9392403364181519,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2238152027130127,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5892173647880554,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.644720196723938,],\n",
+      "[-0.1283257156610489,],\n",
+      "[0.5480666756629944,],\n",
+      "[-0.9999391436576843,],\n",
+      "[0.009961729869246483,],\n",
+      "[0.456155002117157,],\n",
+      "[-0.3495990037918091,],\n",
+      "[0.9875202775001526,],\n",
+      "[0.8271691203117371,],\n",
+      "[-0.9392403364181519,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3342715799808502,],\n",
+      "[-0.13424375653266907,],\n",
+      "[-0.3841778039932251,],\n",
+      "[0.0588737390935421,],\n",
+      "[-0.2877541482448578,],\n",
+      "[-0.018316524103283882,],\n",
+      "[-0.4636456370353699,],\n",
+      "[-0.19539949297904968,],\n",
+      "[-0.04421583563089371,],\n",
+      "[0.005015056114643812,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2238152027130127,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.5352100133895874,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6443859338760376,],\n",
+      "[-0.12819147109985352,],\n",
+      "[0.5484508275985718,],\n",
+      "[-0.9999980330467224,],\n",
+      "[0.010249484330415726,],\n",
+      "[0.4561733305454254,],\n",
+      "[-0.3491353690624237,],\n",
+      "[0.9877156615257263,],\n",
+      "[0.8272133469581604,],\n",
+      "[-0.939245343208313,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.22227999567985535,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.9422525763511658,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6443859338760376,],\n",
+      "[-0.12819147109985352,],\n",
+      "[0.5484508275985718,],\n",
+      "[-0.9999980330467224,],\n",
+      "[0.010249484330415726,],\n",
+      "[0.4561733305454254,],\n",
+      "[-0.3491353690624237,],\n",
+      "[0.9877156615257263,],\n",
+      "[0.8272133469581604,],\n",
+      "[-0.939245343208313,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4342115521430969,],\n",
+      "[-0.16332866251468658,],\n",
+      "[-0.44270220398902893,],\n",
+      "[0.06636527925729752,],\n",
+      "[-0.24084094166755676,],\n",
+      "[-0.014859026297926903,],\n",
+      "[-0.38887184858322144,],\n",
+      "[-0.18670549988746643,],\n",
+      "[-0.03535394370555878,],\n",
+      "[0.005265730433166027,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.22227999567985535,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.9413942098617554,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6439517140388489,],\n",
+      "[-0.12802813947200775,],\n",
+      "[0.5488935112953186,],\n",
+      "[-1.0000643730163574,],\n",
+      "[0.010490325279533863,],\n",
+      "[0.4561882019042969,],\n",
+      "[-0.34874650835990906,],\n",
+      "[0.9879023432731628,],\n",
+      "[0.8272486925125122,],\n",
+      "[-0.9392505884170532,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2203385978937149,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.3663315773010254,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6439517140388489,],\n",
+      "[-0.12802813947200775,],\n",
+      "[0.5488935112953186,],\n",
+      "[-1.0000643730163574,],\n",
+      "[0.010490325279533863,],\n",
+      "[0.4561882019042969,],\n",
+      "[-0.34874650835990906,],\n",
+      "[0.9879023432731628,],\n",
+      "[0.8272486925125122,],\n",
+      "[-0.9392505884170532,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.511968195438385,],\n",
+      "[-0.1823951005935669,],\n",
+      "[-0.47619038820266724,],\n",
+      "[0.07033713907003403,],\n",
+      "[-0.19618652760982513,],\n",
+      "[-0.01152937114238739,],\n",
+      "[-0.3155256509780884,],\n",
+      "[-0.18252354860305786,],\n",
+      "[-0.028133992105722427,],\n",
+      "[0.0056543308310210705,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2203385978937149,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.337803363800049,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.643439769744873,],\n",
+      "[-0.12784574925899506,],\n",
+      "[0.5493696928024292,],\n",
+      "[-1.0001347064971924,],\n",
+      "[0.010686512105166912,],\n",
+      "[0.4561997354030609,],\n",
+      "[-0.3484309911727905,],\n",
+      "[0.9880848526954651,],\n",
+      "[0.8272768259048462,],\n",
+      "[-0.9392562508583069,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21800079941749573,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.326645016670227,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.643439769744873,],\n",
+      "[-0.12784574925899506,],\n",
+      "[0.5493696928024292,],\n",
+      "[-1.0001347064971924,],\n",
+      "[0.010686512105166912,],\n",
+      "[0.4561997354030609,],\n",
+      "[-0.3484309911727905,],\n",
+      "[0.9880848526954651,],\n",
+      "[0.8272768259048462,],\n",
+      "[-0.9392562508583069,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.47118085622787476,],\n",
+      "[-0.1611129641532898,],\n",
+      "[-0.4116482734680176,],\n",
+      "[0.06032214313745499,],\n",
+      "[-0.2020261585712433,],\n",
+      "[-0.011067399755120277,],\n",
+      "[-0.32006537914276123,],\n",
+      "[-0.19444380700588226,],\n",
+      "[-0.03012680634856224,],\n",
+      "[0.006341138388961554,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21800079941749573,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.303600788116455,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6429685950279236,],\n",
+      "[-0.12768463790416718,],\n",
+      "[0.549781322479248,],\n",
+      "[-1.000195026397705,],\n",
+      "[0.0108885383233428,],\n",
+      "[0.45621079206466675,],\n",
+      "[-0.3481109142303467,],\n",
+      "[0.9882792830467224,],\n",
+      "[0.8273069262504578,],\n",
+      "[-0.9392625689506531,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21569719910621643,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.7147471904754639,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6429685950279236,],\n",
+      "[-0.12768463790416718,],\n",
+      "[0.549781322479248,],\n",
+      "[-1.000195026397705,],\n",
+      "[0.0108885383233428,],\n",
+      "[0.45621079206466675,],\n",
+      "[-0.3481109142303467,],\n",
+      "[0.9882792830467224,],\n",
+      "[0.8273069262504578,],\n",
+      "[-0.9392625689506531,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3090350925922394,],\n",
+      "[-0.10276642441749573,],\n",
+      "[-0.2608281075954437,],\n",
+      "[0.03817479684948921,],\n",
+      "[-0.2733570337295532,],\n",
+      "[-0.014132605865597725,],\n",
+      "[-0.4298032224178314,],\n",
+      "[-0.2191188633441925,],\n",
+      "[-0.04166821762919426,],\n",
+      "[0.0071133458986878395,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21569719910621643,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.6908528804779053,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6426595449447632,],\n",
+      "[-0.1275818645954132,],\n",
+      "[0.5500421524047852,],\n",
+      "[-1.0002331733703613,],\n",
+      "[0.011161895468831062,],\n",
+      "[0.4562249183654785,],\n",
+      "[-0.34768110513687134,],\n",
+      "[0.9884983897209167,],\n",
+      "[0.8273485898971558,],\n",
+      "[-0.9392696619033813,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21400634944438934,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1288204789161682,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6426595449447632,],\n",
+      "[-0.1275818645954132,],\n",
+      "[0.5500421524047852,],\n",
+      "[-1.0002331733703613,],\n",
+      "[0.011161895468831062,],\n",
+      "[0.4562249183654785,],\n",
+      "[-0.34768110513687134,],\n",
+      "[0.9884983897209167,],\n",
+      "[0.8273485898971558,],\n",
+      "[-0.9392696619033813,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.112700454890728,],\n",
+      "[-0.036861225962638855,],\n",
+      "[-0.09424661844968796,],\n",
+      "[0.0138649707660079,],\n",
+      "[-0.3830631971359253,],\n",
+      "[-0.01923350617289543,],\n",
+      "[-0.6055594086647034,],\n",
+      "[-0.2449668049812317,],\n",
+      "[-0.05611696094274521,],\n",
+      "[0.007692549377679825,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21400634944438934,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7178310751914978,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425468325614929,],\n",
+      "[-0.12754499912261963,],\n",
+      "[0.550136387348175,],\n",
+      "[-1.0002470016479492,],\n",
+      "[0.01154495868831873,],\n",
+      "[0.4562441408634186,],\n",
+      "[-0.34707555174827576,],\n",
+      "[0.9887433648109436,],\n",
+      "[0.8274046778678894,],\n",
+      "[-0.9392773509025574,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21328851580619812,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.006492191925644875,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425468325614929,],\n",
+      "[-0.12754499912261963,],\n",
+      "[0.550136387348175,],\n",
+      "[-1.0002470016479492,],\n",
+      "[0.01154495868831873,],\n",
+      "[0.4562441408634186,],\n",
+      "[-0.34707555174827576,],\n",
+      "[0.9887433648109436,],\n",
+      "[0.8274046778678894,],\n",
+      "[-0.9392773509025574,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.02101914957165718,],\n",
+      "[0.00682091386988759,],\n",
+      "[0.017794905230402946,],\n",
+      "[-0.0026456695050001144,],\n",
+      "[-0.4795096516609192,],\n",
+      "[-0.02396758459508419,],\n",
+      "[-0.7651053667068481,],\n",
+      "[-0.26188045740127563,],\n",
+      "[-0.0660875141620636,],\n",
+      "[0.007967191748321056,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21328851580619812,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.16114817559719086,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425678730010986,],\n",
+      "[-0.1275518238544464,],\n",
+      "[0.5501185655593872,],\n",
+      "[-1.000244379043579,],\n",
+      "[0.012024468742311,],\n",
+      "[0.4562681019306183,],\n",
+      "[-0.34631043672561646,],\n",
+      "[0.9890052676200867,],\n",
+      "[0.8274707794189453,],\n",
+      "[-0.9392853379249573,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2134496569633484,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07035397738218307,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425678730010986,],\n",
+      "[-0.1275518238544464,],\n",
+      "[0.5501185655593872,],\n",
+      "[-1.000244379043579,],\n",
+      "[0.012024468742311,],\n",
+      "[0.4562681019306183,],\n",
+      "[-0.34631043672561646,],\n",
+      "[0.9890052676200867,],\n",
+      "[0.8274707794189453,],\n",
+      "[-0.9392853379249573,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.055928997695446014,],\n",
+      "[0.01813119277358055,],\n",
+      "[0.048790961503982544,],\n",
+      "[-0.007361642085015774,],\n",
+      "[-0.5185478329658508,],\n",
+      "[-0.026190252974629402,],\n",
+      "[-0.8326623439788818,],\n",
+      "[-0.2666972577571869,],\n",
+      "[-0.06873677670955658,],\n",
+      "[0.008006502874195576,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2134496569633484,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5304867029190063,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.642623782157898,],\n",
+      "[-0.12756995856761932,],\n",
+      "[0.5500697493553162,],\n",
+      "[-1.000236988067627,],\n",
+      "[0.01254301704466343,],\n",
+      "[0.45629429817199707,],\n",
+      "[-0.34547775983810425,],\n",
+      "[0.9892719388008118,],\n",
+      "[0.8275395035743713,],\n",
+      "[-0.9392933249473572,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21398013830184937,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0176565982401371,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.642623782157898,],\n",
+      "[-0.12756995856761932,],\n",
+      "[0.5500697493553162,],\n",
+      "[-1.000236988067627,],\n",
+      "[0.01254301704466343,],\n",
+      "[0.45629429817199707,],\n",
+      "[-0.34547775983810425,],\n",
+      "[0.9892719388008118,],\n",
+      "[0.8275395035743713,],\n",
+      "[-0.9392933249473572,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0221470408141613,],\n",
+      "[0.007218498270958662,],\n",
+      "[0.020196663215756416,],\n",
+      "[-0.0031018415465950966,],\n",
+      "[-0.4899744391441345,],\n",
+      "[-0.02494346722960472,],\n",
+      "[-0.7846900820732117,],\n",
+      "[-0.2628224492073059,],\n",
+      "[-0.06625664234161377,],\n",
+      "[0.007945987395942211,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21398013830184937,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2657563388347626,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6426459550857544,],\n",
+      "[-0.12757717072963715,],\n",
+      "[0.5500495433807373,],\n",
+      "[-1.0002338886260986,],\n",
+      "[0.01303299143910408,],\n",
+      "[0.45631924271583557,],\n",
+      "[-0.34469306468963623,],\n",
+      "[0.9895347356796265,],\n",
+      "[0.8276057839393616,],\n",
+      "[-0.9393012523651123,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21424590051174164,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.037943366914987564,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6426459550857544,],\n",
+      "[-0.12757717072963715,],\n",
+      "[0.5500495433807373,],\n",
+      "[-1.0002338886260986,],\n",
+      "[0.01303299143910408,],\n",
+      "[0.45631924271583557,],\n",
+      "[-0.34469306468963623,],\n",
+      "[0.9895347356796265,],\n",
+      "[0.8276057839393616,],\n",
+      "[-0.9393012523651123,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.025197692215442657,],\n",
+      "[-0.00831346120685339,],\n",
+      "[-0.02430151030421257,],\n",
+      "[0.0038074797485023737,],\n",
+      "[-0.42272433638572693,],\n",
+      "[-0.021235166117548943,],\n",
+      "[-0.6647928953170776,],\n",
+      "[-0.256430059671402,],\n",
+      "[-0.06277459859848022,],\n",
+      "[0.007888534106314182,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21424590051174164,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.38958168029785156,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6426207423210144,],\n",
+      "[-0.12756885588169098,],\n",
+      "[0.5500738620758057,],\n",
+      "[-1.0002377033233643,],\n",
+      "[0.013455715961754322,],\n",
+      "[0.456340491771698,],\n",
+      "[-0.3440282642841339,],\n",
+      "[0.9897911548614502,],\n",
+      "[0.8276685476303101,],\n",
+      "[-0.9393091201782227,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21385632455348969,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.23526206612586975,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6426207423210144,],\n",
+      "[-0.12756885588169098,],\n",
+      "[0.5500738620758057,],\n",
+      "[-1.0002377033233643,],\n",
+      "[0.013455715961754322,],\n",
+      "[0.456340491771698,],\n",
+      "[-0.3440282642841339,],\n",
+      "[0.9897911548614502,],\n",
+      "[0.8276685476303101,],\n",
+      "[-0.9393091201782227,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04798385500907898,],\n",
+      "[-0.01613350585103035,],\n",
+      "[-0.04939036816358566,],\n",
+      "[0.007907952181994915,],\n",
+      "[-0.3661513328552246,],\n",
+      "[-0.017710484564304352,],\n",
+      "[-0.5580495595932007,],\n",
+      "[-0.25239816308021545,],\n",
+      "[-0.06111107021570206,],\n",
+      "[0.007872444577515125,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21385632455348969,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.970077395439148,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425727605819702,],\n",
+      "[-0.12755271792411804,],\n",
+      "[0.5501232743263245,],\n",
+      "[-1.0002455711364746,],\n",
+      "[0.013821867294609547,],\n",
+      "[0.4563581943511963,],\n",
+      "[-0.3434702157974243,],\n",
+      "[0.990043580532074,],\n",
+      "[0.8277296423912048,],\n",
+      "[-0.939316987991333,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.212886244058609,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2855167090892792,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425727605819702,],\n",
+      "[-0.12755271792411804,],\n",
+      "[0.5501232743263245,],\n",
+      "[-1.0002455711364746,],\n",
+      "[0.013821867294609547,],\n",
+      "[0.4563581943511963,],\n",
+      "[-0.3434702157974243,],\n",
+      "[0.990043580532074,],\n",
+      "[0.8277296423912048,],\n",
+      "[-0.939316987991333,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.03996863588690758,],\n",
+      "[-0.013768280856311321,],\n",
+      "[-0.04420439526438713,],\n",
+      "[0.007243116851896048,],\n",
+      "[-0.3592364490032196,],\n",
+      "[-0.016841690987348557,],\n",
+      "[-0.5394473671913147,],\n",
+      "[-0.25229141116142273,],\n",
+      "[-0.06167934089899063,],\n",
+      "[0.00788965169340372,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.212886244058609,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0686753988265991,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425327658653259,],\n",
+      "[-0.12753894925117493,],\n",
+      "[0.5501675009727478,],\n",
+      "[-1.0002528429031372,],\n",
+      "[0.014181103557348251,],\n",
+      "[0.45637503266334534,],\n",
+      "[-0.34293076395988464,],\n",
+      "[0.9902958869934082,],\n",
+      "[0.8277913331985474,],\n",
+      "[-0.9393248558044434,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21181756258010864,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08460670709609985,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425327658653259,],\n",
+      "[-0.12753894925117493,],\n",
+      "[0.5501675009727478,],\n",
+      "[-1.0002528429031372,],\n",
+      "[0.014181103557348251,],\n",
+      "[0.45637503266334534,],\n",
+      "[-0.34293076395988464,],\n",
+      "[0.9902958869934082,],\n",
+      "[0.8277913331985474,],\n",
+      "[-0.9393248558044434,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.016316192224621773,],\n",
+      "[-0.005774335935711861,],\n",
+      "[-0.019476937130093575,],\n",
+      "[0.003269920591264963,],\n",
+      "[-0.40753158926963806,],\n",
+      "[-0.019209139049053192,],\n",
+      "[-0.6282907128334045,],\n",
+      "[-0.25487515330314636,],\n",
+      "[-0.06334583461284637,],\n",
+      "[0.007916311733424664,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21181756258010864,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5817448496818542,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425164341926575,],\n",
+      "[-0.1275331676006317,],\n",
+      "[0.5501869916915894,],\n",
+      "[-1.000256061553955,],\n",
+      "[0.014588635414838791,],\n",
+      "[0.4563942551612854,],\n",
+      "[-0.34230247139930725,],\n",
+      "[0.9905507564544678,],\n",
+      "[0.8278546929359436,],\n",
+      "[-0.9393327832221985,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21123582124710083,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.013208149932324886,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425164341926575,],\n",
+      "[-0.1275331676006317,],\n",
+      "[0.5501869916915894,],\n",
+      "[-1.000256061553955,],\n",
+      "[0.014588635414838791,],\n",
+      "[0.4563942551612854,],\n",
+      "[-0.34230247139930725,],\n",
+      "[0.9905507564544678,],\n",
+      "[0.8278546929359436,],\n",
+      "[-0.9393327832221985,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.004807165823876858,],\n",
+      "[0.0017480218084529042,],\n",
+      "[0.006210696417838335,],\n",
+      "[-0.0010694145457819104,],\n",
+      "[-0.4821101725101471,],\n",
+      "[-0.023398563265800476,],\n",
+      "[-0.7769507765769958,],\n",
+      "[-0.2580000162124634,],\n",
+      "[-0.0648130401968956,],\n",
+      "[0.007934941910207272,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21123582124710083,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.22985339164733887,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425212621688843,],\n",
+      "[-0.1275349110364914,],\n",
+      "[0.5501807928085327,],\n",
+      "[-1.0002549886703491,],\n",
+      "[0.015070745721459389,],\n",
+      "[0.45641764998435974,],\n",
+      "[-0.34152552485466003,],\n",
+      "[0.9908087849617004,],\n",
+      "[0.8279194831848145,],\n",
+      "[-0.9393407106399536,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21146567165851593,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2028287947177887,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425212621688843,],\n",
+      "[-0.1275349110364914,],\n",
+      "[0.5501807928085327,],\n",
+      "[-1.0002549886703491,],\n",
+      "[0.015070745721459389,],\n",
+      "[0.45641764998435974,],\n",
+      "[-0.34152552485466003,],\n",
+      "[0.9908087849617004,],\n",
+      "[0.8279194831848145,],\n",
+      "[-0.9393407106399536,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.013991905376315117,],\n",
+      "[0.00522030983120203,],\n",
+      "[0.0195961594581604,],\n",
+      "[-0.0034635935444384813,],\n",
+      "[-0.5396418571472168,],\n",
+      "[-0.027089359238743782,],\n",
+      "[-0.9003069996833801,],\n",
+      "[-0.2601337134838104,],\n",
+      "[-0.06544121354818344,],\n",
+      "[0.007940540090203285,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21146567165851593,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9007318019866943,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425352692604065,],\n",
+      "[-0.12754012644290924,],\n",
+      "[0.5501611828804016,],\n",
+      "[-1.0002515316009521,],\n",
+      "[0.015610387548804283,],\n",
+      "[0.45644474029541016,],\n",
+      "[-0.34062522649765015,],\n",
+      "[0.9910688996315002,],\n",
+      "[0.8279849290847778,],\n",
+      "[-0.9393486380577087,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21236640214920044,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2729607820510864,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425352692604065,],\n",
+      "[-0.12754012644290924,],\n",
+      "[0.5501611828804016,],\n",
+      "[-1.0002515316009521,],\n",
+      "[0.015610387548804283,],\n",
+      "[0.45644474029541016,],\n",
+      "[-0.34062522649765015,],\n",
+      "[0.9910688996315002,],\n",
+      "[0.8279849290847778,],\n",
+      "[-0.9393486380577087,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.012020116671919823,],\n",
+      "[0.00459558330476284,],\n",
+      "[0.01826876401901245,],\n",
+      "[-0.0033166261855512857,],\n",
+      "[-0.5492091178894043,],\n",
+      "[-0.028257958590984344,],\n",
+      "[-0.9273989796638489,],\n",
+      "[-0.26083269715309143,],\n",
+      "[-0.06530803442001343,],\n",
+      "[0.007937359623610973,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21236640214920044,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0449126958847046,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754471600055695,],\n",
+      "[0.5501429438591003,],\n",
+      "[-1.0002481937408447,],\n",
+      "[0.016159595921635628,],\n",
+      "[0.4564729928970337,],\n",
+      "[-0.33969783782958984,],\n",
+      "[0.9913297295570374,],\n",
+      "[0.8280502557754517,],\n",
+      "[-0.9393565654754639,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21341131627559662,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08745809644460678,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754471600055695,],\n",
+      "[0.5501429438591003,],\n",
+      "[-1.0002481937408447,],\n",
+      "[0.016159595921635628,],\n",
+      "[0.4564729928970337,],\n",
+      "[-0.33969783782958984,],\n",
+      "[0.9913297295570374,],\n",
+      "[0.8280502557754517,],\n",
+      "[-0.9393565654754639,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005026067607104778,],\n",
+      "[0.0019693749491125345,],\n",
+      "[0.008297163993120193,],\n",
+      "[-0.001547845546156168,],\n",
+      "[-0.5097202658653259,],\n",
+      "[-0.026239126920700073,],\n",
+      "[-0.8446283340454102,],\n",
+      "[-0.2604488134384155,],\n",
+      "[-0.06484462320804596,],\n",
+      "[0.007931635715067387,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21341131627559662,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5914664268493652,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425523161888123,],\n",
+      "[-0.12754668295383453,],\n",
+      "[0.5501346588134766,],\n",
+      "[-1.0002466440200806,],\n",
+      "[0.016669316217303276,],\n",
+      "[0.45649921894073486,],\n",
+      "[-0.33885321021080017,],\n",
+      "[0.9915902018547058,],\n",
+      "[0.8281151056289673,],\n",
+      "[-0.939364492893219,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.214002788066864,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.007336239796131849,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425523161888123,],\n",
+      "[-0.12754668295383453,],\n",
+      "[0.5501346588134766,],\n",
+      "[-1.0002466440200806,],\n",
+      "[0.016669316217303276,],\n",
+      "[0.45649921894073486,],\n",
+      "[-0.33885321021080017,],\n",
+      "[0.9915902018547058,],\n",
+      "[0.8281151056289673,],\n",
+      "[-0.939364492893219,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0010730515932664275,],\n",
+      "[-0.0004316760750953108,],\n",
+      "[-0.0019258441170677543,],\n",
+      "[0.00036921611172147095,],\n",
+      "[-0.44884294271469116,],\n",
+      "[-0.022330274805426598,],\n",
+      "[-0.7046301960945129,],\n",
+      "[-0.25957536697387695,],\n",
+      "[-0.06444749236106873,],\n",
+      "[0.007927718572318554,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.214002788066864,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.17130371928215027,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425512433052063,],\n",
+      "[-0.1275462508201599,],\n",
+      "[0.5501365661621094,],\n",
+      "[-1.0002470016479492,],\n",
+      "[0.01711815968155861,],\n",
+      "[0.45652154088020325,],\n",
+      "[-0.3381485939025879,],\n",
+      "[0.9918497800827026,],\n",
+      "[0.8281795382499695,],\n",
+      "[-0.9393724203109741,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21383148431777954,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14823651313781738,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425512433052063,],\n",
+      "[-0.1275462508201599,],\n",
+      "[0.5501365661621094,],\n",
+      "[-1.0002470016479492,],\n",
+      "[0.01711815968155861,],\n",
+      "[0.45652154088020325,],\n",
+      "[-0.3381485939025879,],\n",
+      "[0.9918497800827026,],\n",
+      "[0.8281795382499695,],\n",
+      "[-0.9393724203109741,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0035497299395501614,],\n",
+      "[-0.0014698368031531572,],\n",
+      "[-0.006931926589459181,],\n",
+      "[0.0013656321680173278,],\n",
+      "[-0.4048926830291748,],\n",
+      "[-0.01910972408950329,],\n",
+      "[-0.5945419669151306,],\n",
+      "[-0.2586928606033325,],\n",
+      "[-0.06428919732570648,],\n",
+      "[0.007926490157842636,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21383148431777954,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7700308561325073,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425476670265198,],\n",
+      "[-0.12754477560520172,],\n",
+      "[0.5501434803009033,],\n",
+      "[-1.0002483129501343,],\n",
+      "[0.01752305217087269,],\n",
+      "[0.45654064416885376,],\n",
+      "[-0.33755403757095337,],\n",
+      "[0.9921084642410278,],\n",
+      "[0.8282438516616821,],\n",
+      "[-0.9393803477287292,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21306145191192627,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.17352934181690216,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425476670265198,],\n",
+      "[-0.12754477560520172,],\n",
+      "[0.5501434803009033,],\n",
+      "[-1.0002483129501343,],\n",
+      "[0.01752305217087269,],\n",
+      "[0.45654064416885376,],\n",
+      "[-0.33755403757095337,],\n",
+      "[0.9921084642410278,],\n",
+      "[0.8282438516616821,],\n",
+      "[-0.9393803477287292,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0028231176547706127,],\n",
+      "[-0.0012055085971951485,],\n",
+      "[-0.006001920439302921,],\n",
+      "[0.0012148742098361254,],\n",
+      "[-0.4028899669647217,],\n",
+      "[-0.018628301098942757,],\n",
+      "[-0.582710862159729,],\n",
+      "[-0.25806373357772827,],\n",
+      "[-0.06433557718992233,],\n",
+      "[0.007927187718451023,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21306145191192627,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8331362009048462,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425448656082153,],\n",
+      "[-0.12754356861114502,],\n",
+      "[0.5501495003700256,],\n",
+      "[-1.0002495050430298,],\n",
+      "[0.017925942316651344,],\n",
+      "[0.45655927062034607,],\n",
+      "[-0.33697131276130676,],\n",
+      "[0.9923665523529053,],\n",
+      "[0.8283081650733948,],\n",
+      "[-0.9393882751464844,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21222831308841705,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.024982765316963196,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425448656082153,],\n",
+      "[-0.12754356861114502,],\n",
+      "[0.5501495003700256,],\n",
+      "[-1.0002495050430298,],\n",
+      "[0.017925942316651344,],\n",
+      "[0.45655927062034607,],\n",
+      "[-0.33697131276130676,],\n",
+      "[0.9923665523529053,],\n",
+      "[0.8283081650733948,],\n",
+      "[-0.9393882751464844,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0007868176908232272,],\n",
+      "[-0.0003465722838882357,],\n",
+      "[-0.0018213751027360559,],\n",
+      "[0.0003788076573982835,],\n",
+      "[-0.44045308232307434,],\n",
+      "[-0.021057240664958954,],\n",
+      "[-0.6772504448890686,],\n",
+      "[-0.25782695412635803,],\n",
+      "[-0.06446170806884766,],\n",
+      "[0.007928384467959404,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21222831308841705,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3161189556121826,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425440907478333,],\n",
+      "[-0.12754322588443756,],\n",
+      "[0.5501513481140137,],\n",
+      "[-1.0002498626708984,],\n",
+      "[0.018366394564509392,],\n",
+      "[0.45658034086227417,],\n",
+      "[-0.3362940549850464,],\n",
+      "[0.9926244020462036,],\n",
+      "[0.828372597694397,],\n",
+      "[-0.9393962025642395,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21191219985485077,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.055609747767448425,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425440907478333,],\n",
+      "[-0.12754322588443756,],\n",
+      "[0.5501513481140137,],\n",
+      "[-1.0002498626708984,],\n",
+      "[0.018366394564509392,],\n",
+      "[0.45658034086227417,],\n",
+      "[-0.3362940549850464,],\n",
+      "[0.9926244020462036,],\n",
+      "[0.828372597694397,],\n",
+      "[-0.9393962025642395,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0008619086001999676,],\n",
+      "[0.0003910469531547278,],\n",
+      "[0.002171829342842102,],\n",
+      "[-0.00046426436165347695,],\n",
+      "[-0.49202048778533936,],\n",
+      "[-0.02483082190155983,],\n",
+      "[-0.8214352130889893,],\n",
+      "[-0.2580721080303192,],\n",
+      "[-0.06456153094768524,],\n",
+      "[0.007929174229502678,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21191219985485077,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.47163429856300354,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425449252128601,],\n",
+      "[-0.1275436133146286,],\n",
+      "[0.5501492023468018,],\n",
+      "[-1.0002493858337402,],\n",
+      "[0.0188584141433239,],\n",
+      "[0.4566051661968231,],\n",
+      "[-0.33547261357307434,],\n",
+      "[0.992882490158081,],\n",
+      "[0.8284371495246887,],\n",
+      "[-0.9394041299819946,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21238383650779724,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.27818500995635986,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425449252128601,],\n",
+      "[-0.1275436133146286,],\n",
+      "[0.5501492023468018,],\n",
+      "[-1.0002493858337402,],\n",
+      "[0.0188584141433239,],\n",
+      "[0.4566051661968231,],\n",
+      "[-0.33547261357307434,],\n",
+      "[0.992882490158081,],\n",
+      "[0.8284371495246887,],\n",
+      "[-0.9394041299819946,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0014150312636047602,],\n",
+      "[0.0006598069448955357,],\n",
+      "[0.003879306372255087,],\n",
+      "[-0.0008527584141120315,],\n",
+      "[-0.5264005064964294,],\n",
+      "[-0.027748696506023407,],\n",
+      "[-0.9282218813896179,],\n",
+      "[-0.2587839663028717,],\n",
+      "[-0.06459512561559677,],\n",
+      "[0.007929314859211445,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21238383650779724,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0548654794692993,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425463557243347,],\n",
+      "[-0.12754426896572113,],\n",
+      "[0.5501453280448914,],\n",
+      "[-1.0002485513687134,],\n",
+      "[0.01938481442630291,],\n",
+      "[0.45663291215896606,],\n",
+      "[-0.3345443904399872,],\n",
+      "[0.9931412935256958,],\n",
+      "[0.8285017609596252,],\n",
+      "[-0.9394120573997498,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2134387046098709,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2922840416431427,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425463557243347,],\n",
+      "[-0.12754426896572113,],\n",
+      "[0.5501453280448914,],\n",
+      "[-1.0002485513687134,],\n",
+      "[0.01938481442630291,],\n",
+      "[0.45663291215896606,],\n",
+      "[-0.3345443904399872,],\n",
+      "[0.9931412935256958,],\n",
+      "[0.8285017609596252,],\n",
+      "[-0.9394120573997498,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0010644214926287532,],\n",
+      "[0.0005092552164569497,],\n",
+      "[0.0031735498923808336,],\n",
+      "[-0.0007176636718213558,],\n",
+      "[-0.5251741409301758,],\n",
+      "[-0.028154507279396057,],\n",
+      "[-0.9331499934196472,],\n",
+      "[-0.25973835587501526,],\n",
+      "[-0.06457409262657166,],\n",
+      "[0.007929058745503426,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2134387046098709,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0812652111053467,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425474286079407,],\n",
+      "[-0.12754477560520172,],\n",
+      "[0.5501421689987183,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.019909989088773727,],\n",
+      "[0.45666107535362244,],\n",
+      "[-0.33361124992370605,],\n",
+      "[0.993401050567627,],\n",
+      "[0.828566312789917,],\n",
+      "[-0.9394199848175049,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21451996266841888,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07176020741462708,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425474286079407,],\n",
+      "[-0.12754477560520172,],\n",
+      "[0.5501421689987183,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.019909989088773727,],\n",
+      "[0.45666107535362244,],\n",
+      "[-0.33361124992370605,],\n",
+      "[0.993401050567627,],\n",
+      "[0.828566312789917,],\n",
+      "[-0.9394199848175049,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0003869594365824014,],\n",
+      "[0.00018994478159584105,],\n",
+      "[0.0012546045472845435,],\n",
+      "[-0.00029188403277657926,],\n",
+      "[-0.4920027554035187,],\n",
+      "[-0.025847177952528,],\n",
+      "[-0.8338779807090759,],\n",
+      "[-0.2605496644973755,],\n",
+      "[-0.06453483551740646,],\n",
+      "[0.007928729988634586,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21451996266841888,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5357622504234314,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425477862358093,],\n",
+      "[-0.12754496932029724,],\n",
+      "[0.550140917301178,],\n",
+      "[-1.000247597694397,],\n",
+      "[0.02040199190378189,],\n",
+      "[0.45668691396713257,],\n",
+      "[-0.33277738094329834,],\n",
+      "[0.9936615824699402,],\n",
+      "[0.8286308646202087,],\n",
+      "[-0.93942791223526,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21505571901798248,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.015625247731804848,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425477862358093,],\n",
+      "[-0.12754496932029724,],\n",
+      "[0.550140917301178,],\n",
+      "[-1.000247597694397,],\n",
+      "[0.02040199190378189,],\n",
+      "[0.45668691396713257,],\n",
+      "[-0.33277738094329834,],\n",
+      "[0.9936615824699402,],\n",
+      "[0.8286308646202087,],\n",
+      "[-0.93942791223526,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00013245090667624027,],\n",
+      "[-6.6808242991101e-05,],\n",
+      "[-0.00046713624033145607,],\n",
+      "[0.00011176978296134621,],\n",
+      "[-0.4489539861679077,],\n",
+      "[-0.02231227606534958,],\n",
+      "[-0.6914522647857666,],\n",
+      "[-0.26084035634994507,],\n",
+      "[-0.06450538337230682,],\n",
+      "[0.007928620092570782,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21505571901798248,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.250002384185791,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425476670265198,],\n",
+      "[-0.12754490971565247,],\n",
+      "[0.5501413941383362,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.02085094526410103,],\n",
+      "[0.45670923590660095,],\n",
+      "[-0.3320859372615814,],\n",
+      "[0.9939224123954773,],\n",
+      "[0.8286953568458557,],\n",
+      "[-0.9394358396530151,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2148057222366333,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15955263376235962,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425476670265198,],\n",
+      "[-0.12754490971565247,],\n",
+      "[0.5501413941383362,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.02085094526410103,],\n",
+      "[0.45670923590660095,],\n",
+      "[-0.3320859372615814,],\n",
+      "[0.9939224123954773,],\n",
+      "[0.8286953568458557,],\n",
+      "[-0.9394358396530151,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0003104108909610659,],\n",
+      "[-0.00016124632384162396,],\n",
+      "[-0.001191487186588347,],\n",
+      "[0.00029302926850505173,],\n",
+      "[-0.4221048057079315,],\n",
+      "[-0.019808292388916016,],\n",
+      "[-0.5922305583953857,],\n",
+      "[-0.26048630475997925,],\n",
+      "[-0.06449557840824127,],\n",
+      "[0.007928532548248768,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2148057222366333,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7988811135292053,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473690032959,],\n",
+      "[-0.12754474580287933,],\n",
+      "[0.5501425862312317,],\n",
+      "[-1.0002479553222656,],\n",
+      "[0.02127305045723915,],\n",
+      "[0.45672905445098877,],\n",
+      "[-0.33149370551109314,],\n",
+      "[0.9941828846931458,],\n",
+      "[0.8287598490715027,],\n",
+      "[-0.9394437670707703,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21400684118270874,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1513013243675232,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473690032959,],\n",
+      "[-0.12754474580287933,],\n",
+      "[0.5501425862312317,],\n",
+      "[-1.0002479553222656,],\n",
+      "[0.02127305045723915,],\n",
+      "[0.45672905445098877,],\n",
+      "[-0.33149370551109314,],\n",
+      "[0.9941828846931458,],\n",
+      "[0.8287598490715027,],\n",
+      "[-0.9394437670707703,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0002216681750724092,],\n",
+      "[-0.00011876832286361605,],\n",
+      "[-0.0009263415122404695,],\n",
+      "[0.0002340765786357224,],\n",
+      "[-0.42530199885368347,],\n",
+      "[-0.019820241257548332,],\n",
+      "[-0.5957522988319397,],\n",
+      "[-0.2597268223762512,],\n",
+      "[-0.06450052559375763,],\n",
+      "[0.007928544655442238,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21400684118270874,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7779483199119568,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425471305847168,],\n",
+      "[-0.12754462659358978,],\n",
+      "[0.5501435399055481,],\n",
+      "[-1.0002481937408447,],\n",
+      "[0.021698351949453354,],\n",
+      "[0.4567488729953766,],\n",
+      "[-0.3308979570865631,],\n",
+      "[0.9944425821304321,],\n",
+      "[0.8288243412971497,],\n",
+      "[-0.9394516944885254,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21322889626026154,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.010077179409563541,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425471305847168,],\n",
+      "[-0.12754462659358978,],\n",
+      "[0.5501435399055481,],\n",
+      "[-1.0002481937408447,],\n",
+      "[0.021698351949453354,],\n",
+      "[0.4567488729953766,],\n",
+      "[-0.3308979570865631,],\n",
+      "[0.9944425821304321,],\n",
+      "[0.8288243412971497,],\n",
+      "[-0.9394516944885254,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[4.1949268052121624e-05,],\n",
+      "[-2.3181186406873167e-05,],\n",
+      "[-0.00019084839732386172,],\n",
+      "[4.954930045641959e-05,],\n",
+      "[-0.4530533254146576,],\n",
+      "[-0.022162292152643204,],\n",
+      "[-0.6995219588279724,],\n",
+      "[-0.25905367732048035,],\n",
+      "[-0.06451002508401871,],\n",
+      "[0.007928645238280296,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21322889626026154,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.20077043771743774,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.642547070980072,],\n",
+      "[-0.1275445967912674,],\n",
+      "[0.5501437187194824,],\n",
+      "[-1.0002481937408447,],\n",
+      "[0.02215140499174595,],\n",
+      "[0.45677104592323303,],\n",
+      "[-0.3301984369754791,],\n",
+      "[0.994701623916626,],\n",
+      "[0.8288888335227966,],\n",
+      "[-0.9394596219062805,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21302813291549683,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08429431170225143,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.642547070980072,],\n",
+      "[-0.1275445967912674,],\n",
+      "[0.5501437187194824,],\n",
+      "[-1.0002481937408447,],\n",
+      "[0.02215140499174595,],\n",
+      "[0.45677104592323303,],\n",
+      "[-0.3301984369754791,],\n",
+      "[0.994701623916626,],\n",
+      "[0.8288888335227966,],\n",
+      "[-0.9394596219062805,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.896458166418597e-05,],\n",
+      "[5.06214055349119e-05,],\n",
+      "[0.0004404282080940902,],\n",
+      "[-0.00011753596481867135,],\n",
+      "[-0.4863208532333374,],\n",
+      "[-0.025304704904556274,],\n",
+      "[-0.839996874332428,],\n",
+      "[-0.2589424252510071,],\n",
+      "[-0.06451669335365295,],\n",
+      "[0.007928646169602871,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21302813291549683,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5806689858436584,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425471305847168,],\n",
+      "[-0.12754464149475098,],\n",
+      "[0.550143301486969,],\n",
+      "[-1.0002480745315552,],\n",
+      "[0.022637726739048958,],\n",
+      "[0.4567963480949402,],\n",
+      "[-0.32935842871665955,],\n",
+      "[0.9949605464935303,],\n",
+      "[0.8289533257484436,],\n",
+      "[-0.9394675493240356,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21360880136489868,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.29686304926872253,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425471305847168,],\n",
+      "[-0.12754464149475098,],\n",
+      "[0.550143301486969,],\n",
+      "[-1.0002480745315552,],\n",
+      "[0.022637726739048958,],\n",
+      "[0.4567963480949402,],\n",
+      "[-0.32935842871665955,],\n",
+      "[0.9949605464935303,],\n",
+      "[0.8289533257484436,],\n",
+      "[-0.9394675493240356,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00012242018419783562,],\n",
+      "[7.156923675211146e-05,],\n",
+      "[0.0006590680568479002,],\n",
+      "[-0.0001808976667234674,],\n",
+      "[-0.5052533149719238,],\n",
+      "[-0.02739030495285988,],\n",
+      "[-0.9307806491851807,],\n",
+      "[-0.25955265760421753,],\n",
+      "[-0.064519003033638,],\n",
+      "[0.007928643375635147,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21360880136489868,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0897029638290405,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425472497940063,],\n",
+      "[-0.12754471600055695,],\n",
+      "[0.5501426458358765,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02314298041164875,],\n",
+      "[0.4568237364292145,],\n",
+      "[-0.32842764258384705,],\n",
+      "[0.9952201247215271,],\n",
+      "[0.8290178179740906,],\n",
+      "[-0.9394754767417908,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21469850838184357,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2573834955692291,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425472497940063,],\n",
+      "[-0.12754471600055695,],\n",
+      "[0.5501426458358765,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02314298041164875,],\n",
+      "[0.4568237364292145,],\n",
+      "[-0.32842764258384705,],\n",
+      "[0.9952201247215271,],\n",
+      "[0.8290178179740906,],\n",
+      "[-0.9394754767417908,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.35825121612288e-05,],\n",
+      "[5.013282861909829e-05,],\n",
+      "[0.0004891306161880493,],\n",
+      "[-0.0001381360343657434,],\n",
+      "[-0.5002724528312683,],\n",
+      "[-0.02724887803196907,],\n",
+      "[-0.9164143800735474,],\n",
+      "[-0.2606181502342224,],\n",
+      "[-0.06451686471700668,],\n",
+      "[0.007928639650344849,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21469850838184357,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0146605968475342,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754476070404053,],\n",
+      "[0.5501421689987183,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.0236432533711195,],\n",
+      "[0.45685097575187683,],\n",
+      "[-0.3275112211704254,],\n",
+      "[0.9954807162284851,],\n",
+      "[0.8290823101997375,],\n",
+      "[-0.9394834041595459,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21571317315101624,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.039677560329437256,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754476070404053,],\n",
+      "[0.5501421689987183,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.0236432533711195,],\n",
+      "[0.45685097575187683,],\n",
+      "[-0.3275112211704254,],\n",
+      "[0.9954807162284851,],\n",
+      "[0.8290823101997375,],\n",
+      "[-0.9394834041595459,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.4062263037194498e-05,],\n",
+      "[1.481030722061405e-05,],\n",
+      "[0.00015306967543438077,],\n",
+      "[-4.447608080226928e-05,],\n",
+      "[-0.4763990342617035,],\n",
+      "[-0.02504293993115425,],\n",
+      "[-0.8070672154426575,],\n",
+      "[-0.2615848481655121,],\n",
+      "[-0.06451395153999329,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21571317315101624,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.39838463068008423,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754477560520172,],\n",
+      "[0.5501419901847839,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.024119652807712555,],\n",
+      "[0.4568760097026825,],\n",
+      "[-0.32670414447784424,],\n",
+      "[0.9957423210144043,],\n",
+      "[0.8291468024253845,],\n",
+      "[-0.939491331577301,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21611155569553375,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03758269548416138,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754477560520172,],\n",
+      "[0.5501419901847839,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.024119652807712555,],\n",
+      "[0.4568760097026825,],\n",
+      "[-0.32670414447784424,],\n",
+      "[0.9957423210144043,],\n",
+      "[0.8291468024253845,],\n",
+      "[-0.939491331577301,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.7170694263768382e-05,],\n",
+      "[-1.086388419935247e-05,],\n",
+      "[-0.00011879160592798144,],\n",
+      "[3.549444954842329e-05,],\n",
+      "[-0.44923195242881775,],\n",
+      "[-0.022196736186742783,],\n",
+      "[-0.669945240020752,],\n",
+      "[-0.26194441318511963,],\n",
+      "[-0.064511738717556,],\n",
+      "[0.007928636856377125,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21611155569553375,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.38772526383399963,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754476070404053,],\n",
+      "[0.5501421093940735,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.024568885564804077,],\n",
+      "[0.45689821243286133,],\n",
+      "[-0.32603418827056885,],\n",
+      "[0.9960042834281921,],\n",
+      "[0.8292112946510315,],\n",
+      "[-0.9394992589950562,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21572382748126984,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1888597309589386,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754476070404053,],\n",
+      "[0.5501421093940735,],\n",
+      "[-1.0002477169036865,],\n",
+      "[0.024568885564804077,],\n",
+      "[0.45689821243286133,],\n",
+      "[-0.32603418827056885,],\n",
+      "[0.9960042834281921,],\n",
+      "[0.8292112946510315,],\n",
+      "[-0.9394992589950562,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.8221800675964914e-05,],\n",
+      "[-1.8395434381091036e-05,],\n",
+      "[-0.00021246921096462756,],\n",
+      "[6.524180207634345e-05,],\n",
+      "[-0.4348396360874176,],\n",
+      "[-0.020476609468460083,],\n",
+      "[-0.5863901376724243,],\n",
+      "[-0.2615448236465454,],\n",
+      "[-0.06451097875833511,],\n",
+      "[0.007928636856377125,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21572382748126984,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8691595792770386,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754474580287933,],\n",
+      "[0.5501423478126526,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.025003725662827492,],\n",
+      "[0.4569186866283417,],\n",
+      "[-0.32544779777526855,],\n",
+      "[0.9962658286094666,],\n",
+      "[0.8292757868766785,],\n",
+      "[-0.9395071864128113,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2148546725511551,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14301587641239166,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754474580287933,],\n",
+      "[0.5501423478126526,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.025003725662827492,],\n",
+      "[0.4569186866283417,],\n",
+      "[-0.32544779777526855,],\n",
+      "[0.9962658286094666,],\n",
+      "[0.8292757868766785,],\n",
+      "[-0.9395071864128113,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.8006352547672577e-05,],\n",
+      "[-1.2107711881981231e-05,],\n",
+      "[-0.00014757535245735198,],\n",
+      "[4.6550492697861046e-05,],\n",
+      "[-0.43987786769866943,],\n",
+      "[-0.020827241241931915,],\n",
+      "[-0.6053171753883362,],\n",
+      "[-0.2606847286224365,],\n",
+      "[-0.06451170146465302,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2148546725511551,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.756348729133606,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501424670219421,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.025443604215979576,],\n",
+      "[0.4569395184516907,],\n",
+      "[-0.3248424828052521,],\n",
+      "[0.9965265393257141,],\n",
+      "[0.8293402791023254,],\n",
+      "[-0.9395151138305664,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21409831941127777,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.003647831967100501,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501424670219421,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.025443604215979576,],\n",
+      "[0.4569395184516907,],\n",
+      "[-0.3248424828052521,],\n",
+      "[0.9965265393257141,],\n",
+      "[0.8293402791023254,],\n",
+      "[-0.9395151138305664,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.108467015204951e-06,],\n",
+      "[-1.4620894717154442e-06,],\n",
+      "[-1.880998934211675e-05,],\n",
+      "[6.095601293054642e-06,],\n",
+      "[-0.45877742767333984,],\n",
+      "[-0.02285999432206154,],\n",
+      "[-0.7151886820793152,],\n",
+      "[-0.259946346282959,],\n",
+      "[-0.06451243162155151,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21409831941127777,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.1207946166396141,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501424670219421,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.025902381166815758,],\n",
+      "[0.45696237683296204,],\n",
+      "[-0.32412728667259216,],\n",
+      "[0.9967864751815796,],\n",
+      "[0.8294047713279724,],\n",
+      "[-0.9395230412483215,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21397753059864044,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10496872663497925,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501424670219421,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.025902381166815758,],\n",
+      "[0.45696237683296204,],\n",
+      "[-0.32412728667259216,],\n",
+      "[0.9967864751815796,],\n",
+      "[0.8294047713279724,],\n",
+      "[-0.9395230412483215,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.292682650790084e-06,],\n",
+      "[5.9197186601522844e-06,],\n",
+      "[8.048593736020848e-05,],\n",
+      "[-2.6809708288055845e-05,],\n",
+      "[-0.4788207411766052,],\n",
+      "[-0.025250794366002083,],\n",
+      "[-0.8481634855270386,],\n",
+      "[-0.2598404586315155,],\n",
+      "[-0.0645131766796112,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21397753059864044,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6479777693748474,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501424074172974,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.026381202042102814,],\n",
+      "[0.4569876194000244,],\n",
+      "[-0.32327911257743835,],\n",
+      "[0.9970462918281555,],\n",
+      "[0.8294692635536194,],\n",
+      "[-0.9395309686660767,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2146255075931549,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.29114773869514465,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501424074172974,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.026381202042102814,],\n",
+      "[0.4569876194000244,],\n",
+      "[-0.32327911257743835,],\n",
+      "[0.9970462918281555,],\n",
+      "[0.8294692635536194,],\n",
+      "[-0.9395309686660767,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.0126009328814689e-05,],\n",
+      "[7.425206604239065e-06,],\n",
+      "[0.00010685254528652877,],\n",
+      "[-3.660821676021442e-05,],\n",
+      "[-0.4883423149585724,],\n",
+      "[-0.02658732607960701,],\n",
+      "[-0.9209031462669373,],\n",
+      "[-0.26049137115478516,],\n",
+      "[-0.0645131841301918,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2146255075931549,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0791617631912231,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02686954475939274,],\n",
+      "[0.45701420307159424,],\n",
+      "[-0.32235822081565857,],\n",
+      "[0.997306764125824,],\n",
+      "[0.8295337557792664,],\n",
+      "[-0.9395388960838318,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21570466458797455,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20569248497486115,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02686954475939274,],\n",
+      "[0.45701420307159424,],\n",
+      "[-0.32235822081565857,],\n",
+      "[0.997306764125824,],\n",
+      "[0.8295337557792664,],\n",
+      "[-0.9395388960838318,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.240315087779891e-06,],\n",
+      "[4.694769813795574e-06,],\n",
+      "[7.156728679547086e-05,],\n",
+      "[-2.5228237063856795e-05,],\n",
+      "[-0.48296046257019043,],\n",
+      "[-0.026148289442062378,],\n",
+      "[-0.8892936706542969,],\n",
+      "[-0.2615645229816437,],\n",
+      "[-0.06451244652271271,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21570466458797455,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9070678949356079,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.550142228603363,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.027352504432201385,],\n",
+      "[0.45704033970832825,],\n",
+      "[-0.32146891951560974,],\n",
+      "[0.9975683093070984,],\n",
+      "[0.8295982480049133,],\n",
+      "[-0.9395468235015869,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21661172807216644,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.013500050641596317,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.550142228603363,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.027352504432201385,],\n",
+      "[0.45704033970832825,],\n",
+      "[-0.32146891951560974,],\n",
+      "[0.9975683093070984,],\n",
+      "[0.8295982480049133,],\n",
+      "[-0.9395468235015869,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.1721377859430504e-06,],\n",
+      "[9.051703955265111e-07,],\n",
+      "[1.4611918231821619e-05,],\n",
+      "[-5.2990021686127875e-06,],\n",
+      "[-0.46725139021873474,],\n",
+      "[-0.024287186563014984,],\n",
+      "[-0.7759907841682434,],\n",
+      "[-0.26246559619903564,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21661172807216644,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.23237916827201843,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.550142228603363,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.027819756418466568,],\n",
+      "[0.4570646286010742,],\n",
+      "[-0.32069292664527893,],\n",
+      "[0.9978307485580444,],\n",
+      "[0.8296627402305603,],\n",
+      "[-0.939554750919342,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2168441116809845,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07270768284797668,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.550142228603363,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.027819756418466568,],\n",
+      "[0.4570646286010742,],\n",
+      "[-0.32069292664527893,],\n",
+      "[0.9978307485580444,],\n",
+      "[0.8296627402305603,],\n",
+      "[-0.939554750919342,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.99441069526074e-06,],\n",
+      "[-1.5839192428757087e-06,],\n",
+      "[-2.704000689845998e-05,],\n",
+      "[1.0082328117277939e-05,],\n",
+      "[-0.4513542950153351,],\n",
+      "[-0.022198805585503578,],\n",
+      "[-0.6502780914306641,],\n",
+      "[-0.26269200444221497,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2168441116809845,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5392875671386719,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.550142228603363,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.028271110728383064,],\n",
+      "[0.45708683133125305,],\n",
+      "[-0.3200426399707794,],\n",
+      "[0.9980934262275696,],\n",
+      "[0.8297272324562073,],\n",
+      "[-0.9395626783370972,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21630482375621796,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22244518995285034,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.550142228603363,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.028271110728383064,],\n",
+      "[0.45708683133125305,],\n",
+      "[-0.3200426399707794,],\n",
+      "[0.9980934262275696,],\n",
+      "[0.8297272324562073,],\n",
+      "[-0.9395626783370972,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.5576889584044693e-06,],\n",
+      "[-2.09351014746062e-06,],\n",
+      "[-3.773860225919634e-05,],\n",
+      "[1.4458424629992805e-05,],\n",
+      "[-0.444433331489563,],\n",
+      "[-0.021142147481441498,],\n",
+      "[-0.5843324661254883,],\n",
+      "[-0.26214975118637085,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21630482375621796,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.943280816078186,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02871554344892502,],\n",
+      "[0.4571079611778259,],\n",
+      "[-0.3194583058357239,],\n",
+      "[0.9983555674552917,],\n",
+      "[0.8297917246818542,],\n",
+      "[-0.9395706057548523,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21536153554916382,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1360320895910263,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02871554344892502,],\n",
+      "[0.4571079611778259,],\n",
+      "[-0.3194583058357239,],\n",
+      "[0.9983555674552917,],\n",
+      "[0.8297917246818542,],\n",
+      "[-0.9395706057548523,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.466457206333871e-06,],\n",
+      "[-1.2384965657474822e-06,],\n",
+      "[-2.3555960069643334e-05,],\n",
+      "[9.269714610127266e-06,],\n",
+      "[-0.4492360055446625,],\n",
+      "[-0.021646402776241302,],\n",
+      "[-0.6163256764411926,],\n",
+      "[-0.2612094581127167,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21536153554916382,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7376496195793152,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02916477993130684,],\n",
+      "[0.4571295976638794,],\n",
+      "[-0.31884199380874634,],\n",
+      "[0.9986167550086975,],\n",
+      "[0.8298562169075012,],\n",
+      "[-0.9395785331726074,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21462388336658478,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0006448140484280884,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.02916477993130684,],\n",
+      "[0.4571295976638794,],\n",
+      "[-0.31884199380874634,],\n",
+      "[0.9986167550086975,],\n",
+      "[0.8298562169075012,],\n",
+      "[-0.9395785331726074,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.402503854336828e-08,],\n",
+      "[-6.446889955213919e-08,],\n",
+      "[-1.294278263230808e-06,],\n",
+      "[5.23235428318003e-07,],\n",
+      "[-0.4613274335861206,],\n",
+      "[-0.023264726623892784,],\n",
+      "[-0.728098452091217,],\n",
+      "[-0.2604748010635376,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21462388336658478,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.05078641697764397,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.029626106843352318,],\n",
+      "[0.4571528732776642,],\n",
+      "[-0.31811389327049255,],\n",
+      "[0.998877227306366,],\n",
+      "[0.8299207091331482,],\n",
+      "[-0.9395864605903625,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21457310020923615,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.12106705456972122,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.029626106843352318,],\n",
+      "[0.4571528732776642,],\n",
+      "[-0.31811389327049255,],\n",
+      "[0.998877227306366,],\n",
+      "[0.8299207091331482,],\n",
+      "[-0.9395864605903625,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.436810278704797e-07,],\n",
+      "[6.666064678029215e-07,],\n",
+      "[1.4144845408736728e-05,],\n",
+      "[-5.877972398593556e-06,],\n",
+      "[-0.47279348969459534,],\n",
+      "[-0.024953195825219154,],\n",
+      "[-0.84987872838974,],\n",
+      "[-0.26042699813842773,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21457310020923615,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6958941221237183,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.030098900198936462,],\n",
+      "[0.4571778178215027,],\n",
+      "[-0.31726402044296265,],\n",
+      "[0.9991376399993896,],\n",
+      "[0.8299852013587952,],\n",
+      "[-0.9395943880081177,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21526899933815002,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2728266417980194,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.030098900198936462,],\n",
+      "[0.4571778178215027,],\n",
+      "[-0.31726402044296265,],\n",
+      "[0.9991376399993896,],\n",
+      "[0.8299852013587952,],\n",
+      "[-0.9395943880081177,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.18518458345352e-07,],\n",
+      "[7.535556960647227e-07,],\n",
+      "[1.6924952433328144e-05,],\n",
+      "[-7.234210897877347e-06,],\n",
+      "[-0.4771505892276764,],\n",
+      "[-0.02572442777454853,],\n",
+      "[-0.9036308526992798,],\n",
+      "[-0.2611228823661804,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21526899933815002,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.044655203819275,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.030576050281524658,],\n",
+      "[0.45720353722572327,],\n",
+      "[-0.31636038422584534,],\n",
+      "[0.9993987679481506,],\n",
+      "[0.8300496935844421,],\n",
+      "[-0.9396023154258728,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2163136601448059,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1521473079919815,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.030576050281524658,],\n",
+      "[0.45720353722572327,],\n",
+      "[-0.31636038422584534,],\n",
+      "[0.9993987679481506,],\n",
+      "[0.8300496935844421,],\n",
+      "[-0.9396023154258728,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-4.4815638489126286e-07,],\n",
+      "[4.233434083289467e-07,],\n",
+      "[1.0071262295241468e-05,],\n",
+      "[-4.429072305356385e-06,],\n",
+      "[-0.4727107286453247,],\n",
+      "[-0.02519235759973526,],\n",
+      "[-0.8582815527915955,],\n",
+      "[-0.26216456294059753,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2163136601448059,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.780121922492981,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.031048761680722237,],\n",
+      "[0.45722872018814087,],\n",
+      "[-0.3155021071434021,],\n",
+      "[0.9996609091758728,],\n",
+      "[0.8301141858100891,],\n",
+      "[-0.9396102428436279,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21709378063678741,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0008749460685066879,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.031048761680722237,],\n",
+      "[0.45722872018814087,],\n",
+      "[-0.3155021071434021,],\n",
+      "[0.9996609091758728,],\n",
+      "[0.8301141858100891,],\n",
+      "[-0.9396102428436279,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.4917271801427887e-08,],\n",
+      "[2.416898681190105e-08,],\n",
+      "[6.087056476644648e-07,],\n",
+      "[-2.7536094648894505e-07,],\n",
+      "[-0.46298280358314514,],\n",
+      "[-0.023757731541991234,],\n",
+      "[-0.7470217347145081,],\n",
+      "[-0.26294466853141785,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21709378063678741,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.05915902182459831,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.03151174634695053,],\n",
+      "[0.45725247263908386,],\n",
+      "[-0.31475508213043213,],\n",
+      "[0.999923825263977,],\n",
+      "[0.8301786780357361,],\n",
+      "[-0.9396181702613831,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.217152938246727,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11730096489191055,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.03151174634695053,],\n",
+      "[0.45725247263908386,],\n",
+      "[-0.31475508213043213,],\n",
+      "[0.999923825263977,],\n",
+      "[0.8301786780357361,],\n",
+      "[-0.9396181702613831,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.1153007878638164e-07,],\n",
+      "[-2.1110943748681166e-07,],\n",
+      "[-5.620816864393419e-06,],\n",
+      "[2.6139255169255193e-06,],\n",
+      "[-0.45417192578315735,],\n",
+      "[-0.02233201079070568,],\n",
+      "[-0.6364914774894714,],\n",
+      "[-0.2630038261413574,],\n",
+      "[-0.0645124539732933,],\n",
+      "[0.0079286377876997,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.217152938246727,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6849842667579651,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.031965918838977814,],\n",
+      "[0.45727479457855225,],\n",
+      "[-0.3141185939311981,],\n",
+      "[1.000186800956726,],\n",
+      "[0.8302431702613831,],\n",
+      "[-0.9396260976791382,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.216467946767807,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [7/10], Loss: 0.1173\n",
+      "loss_antes tensor([0.5608254075050354,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6425473093986511,],\n",
+      "[-0.12754473090171814,],\n",
+      "[0.5501422882080078,],\n",
+      "[-1.000247836112976,],\n",
+      "[0.031965918838977814,],\n",
+      "[0.45727479457855225,],\n",
+      "[-0.3141185939311981,],\n",
+      "[1.000186800956726,],\n",
+      "[0.8302431702613831,],\n",
+      "[-0.9396260976791382,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.30154579877853394,],\n",
+      "[-0.1326947957277298,],\n",
+      "[-0.40419769287109375,],\n",
+      "[0.06290814280509949,],\n",
+      "[-0.28919219970703125,],\n",
+      "[-0.015198981389403343,],\n",
+      "[-0.4681627154350281,],\n",
+      "[-0.1850784569978714,],\n",
+      "[-0.04408370703458786,],\n",
+      "[0.005529502406716347,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.216467946767807,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4977662563323975,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6422457695007324,],\n",
+      "[-0.12741203606128693,],\n",
+      "[0.5505464673042297,],\n",
+      "[-1.0003107786178589,],\n",
+      "[0.032255109399557114,],\n",
+      "[0.45728999376296997,],\n",
+      "[-0.3136504292488098,],\n",
+      "[1.0003719329833984,],\n",
+      "[0.8302872776985168,],\n",
+      "[-0.9396316409111023,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2149701863527298,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.547595739364624,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6422457695007324,],\n",
+      "[-0.12741203606128693,],\n",
+      "[0.5505464673042297,],\n",
+      "[-1.0003107786178589,],\n",
+      "[0.032255109399557114,],\n",
+      "[0.45728999376296997,],\n",
+      "[-0.3136504292488098,],\n",
+      "[1.0003719329833984,],\n",
+      "[0.8302872776985168,],\n",
+      "[-0.9396316409111023,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3205249011516571,],\n",
+      "[-0.13113169372081757,],\n",
+      "[-0.3717849552631378,],\n",
+      "[0.056175511330366135,],\n",
+      "[-0.29079169034957886,],\n",
+      "[-0.014609251171350479,],\n",
+      "[-0.46996235847473145,],\n",
+      "[-0.19345462322235107,],\n",
+      "[-0.04141383618116379,],\n",
+      "[0.006289946846663952,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2149701863527298,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4799944162368774,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6419252157211304,],\n",
+      "[-0.12728090584278107,],\n",
+      "[0.5509182810783386,],\n",
+      "[-1.0003669261932373,],\n",
+      "[0.03254590183496475,],\n",
+      "[0.45730459690093994,],\n",
+      "[-0.31318047642707825,],\n",
+      "[1.0005654096603394,],\n",
+      "[0.8303287029266357,],\n",
+      "[-0.9396379590034485,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21349018812179565,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8901832103729248,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6419252157211304,],\n",
+      "[-0.12728090584278107,],\n",
+      "[0.5509182810783386,],\n",
+      "[-1.0003669261932373,],\n",
+      "[0.03254590183496475,],\n",
+      "[0.45730459690093994,],\n",
+      "[-0.31318047642707825,],\n",
+      "[1.0005654096603394,],\n",
+      "[0.8303287029266357,],\n",
+      "[-0.9396379590034485,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4195728302001953,],\n",
+      "[-0.16079682111740112,],\n",
+      "[-0.4315950572490692,],\n",
+      "[0.06384037435054779,],\n",
+      "[-0.24408473074436188,],\n",
+      "[-0.011806673370301723,],\n",
+      "[-0.3950076103210449,],\n",
+      "[-0.18474669754505157,],\n",
+      "[-0.03327023237943649,],\n",
+      "[0.006518981419503689,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21349018812179565,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8869900703430176,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6415056586265564,],\n",
+      "[-0.12712010741233826,],\n",
+      "[0.5513498783111572,],\n",
+      "[-1.0004308223724365,],\n",
+      "[0.032789986580610275,],\n",
+      "[0.45731639862060547,],\n",
+      "[-0.31278547644615173,],\n",
+      "[1.000750184059143,],\n",
+      "[0.8303619623184204,],\n",
+      "[-0.939644455909729,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.21160319447517395,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.3047960996627808,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6415056586265564,],\n",
+      "[-0.12712010741233826,],\n",
+      "[0.5513498783111572,],\n",
+      "[-1.0004308223724365,],\n",
+      "[0.032789986580610275,],\n",
+      "[0.45731639862060547,],\n",
+      "[-0.31278547644615173,],\n",
+      "[1.000750184059143,],\n",
+      "[0.8303619623184204,],\n",
+      "[-0.939644455909729,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.49695131182670593,],\n",
+      "[-0.18044869601726532,],\n",
+      "[-0.4662984311580658,],\n",
+      "[0.068019337952137,],\n",
+      "[-0.19948342442512512,],\n",
+      "[-0.009111860767006874,],\n",
+      "[-0.32144784927368164,],\n",
+      "[-0.18039602041244507,],\n",
+      "[-0.026579923927783966,],\n",
+      "[0.00692220963537693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.21160319447517395,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.284552812576294,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.641008734703064,],\n",
+      "[-0.12693965435028076,],\n",
+      "[0.5518161654472351,],\n",
+      "[-1.00049889087677,],\n",
+      "[0.032989468425512314,],\n",
+      "[0.4573255181312561,],\n",
+      "[-0.31246402859687805,],\n",
+      "[1.0009305477142334,],\n",
+      "[0.8303885459899902,],\n",
+      "[-0.939651370048523,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2093186378479004,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2677077054977417,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.641008734703064,],\n",
+      "[-0.12693965435028076,],\n",
+      "[0.5518161654472351,],\n",
+      "[-1.00049889087677,],\n",
+      "[0.032989468425512314,],\n",
+      "[0.4573255181312561,],\n",
+      "[-0.31246402859687805,],\n",
+      "[1.0009305477142334,],\n",
+      "[0.8303885459899902,],\n",
+      "[-0.939651370048523,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.45698004961013794,],\n",
+      "[-0.15935257077217102,],\n",
+      "[-0.40279626846313477,],\n",
+      "[0.058346159756183624,],\n",
+      "[-0.20483434200286865,],\n",
+      "[-0.00861310213804245,],\n",
+      "[-0.325824111700058,],\n",
+      "[-0.1916038990020752,],\n",
+      "[-0.02828604355454445,],\n",
+      "[0.007692587561905384,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2093186378479004,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.251849889755249,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.640551745891571,],\n",
+      "[-0.12678030133247375,],\n",
+      "[0.5522189736366272,],\n",
+      "[-1.0005571842193604,],\n",
+      "[0.03319430351257324,],\n",
+      "[0.45733413100242615,],\n",
+      "[-0.3121381998062134,],\n",
+      "[1.0011221170425415,],\n",
+      "[0.8304168581962585,],\n",
+      "[-0.939659059047699,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20706678926944733,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6731455326080322,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.640551745891571,],\n",
+      "[-0.12678030133247375,],\n",
+      "[0.5522189736366272,],\n",
+      "[-1.0005571842193604,],\n",
+      "[0.03319430351257324,],\n",
+      "[0.45733413100242615,],\n",
+      "[-0.3121381998062134,],\n",
+      "[1.0011221170425415,],\n",
+      "[0.8304168581962585,],\n",
+      "[-0.939659059047699,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.2971467971801758,],\n",
+      "[-0.10084416717290878,],\n",
+      "[-0.25308576226234436,],\n",
+      "[0.036652397364377975,],\n",
+      "[-0.2751530408859253,],\n",
+      "[-0.010740222409367561,],\n",
+      "[-0.43542569875717163,],\n",
+      "[-0.21513888239860535,],\n",
+      "[-0.03870658203959465,],\n",
+      "[0.008576375432312489,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20706678926944733,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.6409074068069458,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6402546167373657,],\n",
+      "[-0.1266794502735138,],\n",
+      "[0.5524720549583435,],\n",
+      "[-1.0005937814712524,],\n",
+      "[0.03346945717930794,],\n",
+      "[0.4573448598384857,],\n",
+      "[-0.31170278787612915,],\n",
+      "[1.0013372898101807,],\n",
+      "[0.8304555416107178,],\n",
+      "[-0.9396676421165466,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20542588829994202,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11220359802246094,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6402546167373657,],\n",
+      "[-0.1266794502735138,],\n",
+      "[0.5524720549583435,],\n",
+      "[-1.0005937814712524,],\n",
+      "[0.03346945717930794,],\n",
+      "[0.4573448598384857,],\n",
+      "[-0.31170278787612915,],\n",
+      "[1.0013372898101807,],\n",
+      "[0.8304555416107178,],\n",
+      "[-0.9396676421165466,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.10405726730823517,],\n",
+      "[-0.03476313501596451,],\n",
+      "[-0.08784232288599014,],\n",
+      "[0.012800282798707485,],\n",
+      "[-0.3836789131164551,],\n",
+      "[-0.014344287104904652,],\n",
+      "[-0.6109833717346191,],\n",
+      "[-0.2398398071527481,],\n",
+      "[-0.05178738012909889,],\n",
+      "[0.009243322536349297,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20542588829994202,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6699362397193909,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401505470275879,],\n",
+      "[-0.12664468586444855,],\n",
+      "[0.5525599122047424,],\n",
+      "[-1.0006065368652344,],\n",
+      "[0.033853136003017426,],\n",
+      "[0.4573591947555542,],\n",
+      "[-0.311091810464859,],\n",
+      "[1.0015771389007568,],\n",
+      "[0.8305073380470276,],\n",
+      "[-0.9396768808364868,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20475594699382782,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.01069589052349329,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401505470275879,],\n",
+      "[-0.12664468586444855,],\n",
+      "[0.5525599122047424,],\n",
+      "[-1.0006065368652344,],\n",
+      "[0.033853136003017426,],\n",
+      "[0.4573591947555542,],\n",
+      "[-0.311091810464859,],\n",
+      "[1.0015771389007568,],\n",
+      "[0.8305073380470276,],\n",
+      "[-0.9396768808364868,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.026649612933397293,],\n",
+      "[0.008841327391564846,],\n",
+      "[0.022783638909459114,],\n",
+      "[-0.003358763176947832,],\n",
+      "[-0.4791840612888336,],\n",
+      "[-0.01773231290280819,],\n",
+      "[-0.7702085971832275,],\n",
+      "[-0.25598061084747314,],\n",
+      "[-0.06079734489321709,],\n",
+      "[0.009559578262269497,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20475594699382782,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2068418711423874,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401771903038025,],\n",
+      "[-0.1266535222530365,],\n",
+      "[0.5525371432304382,],\n",
+      "[-1.000603199005127,],\n",
+      "[0.03433232009410858,],\n",
+      "[0.4573769271373749,],\n",
+      "[-0.3103215992450714,],\n",
+      "[1.0018330812454224,],\n",
+      "[0.8305681347846985,],\n",
+      "[-0.9396864175796509,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20496279001235962,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08234640210866928,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401771903038025,],\n",
+      "[-0.1266535222530365,],\n",
+      "[0.5525371432304382,],\n",
+      "[-1.000603199005127,],\n",
+      "[0.03433232009410858,],\n",
+      "[0.4573769271373749,],\n",
+      "[-0.3103215992450714,],\n",
+      "[1.0018330812454224,],\n",
+      "[0.8305681347846985,],\n",
+      "[-0.9396864175796509,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.059676192700862885,],\n",
+      "[0.019797179847955704,],\n",
+      "[0.05259178951382637,],\n",
+      "[-0.007876335643231869,],\n",
+      "[-0.517728865146637,],\n",
+      "[-0.019402721896767616,],\n",
+      "[-0.8373596668243408,],\n",
+      "[-0.2605092227458954,],\n",
+      "[-0.06313716620206833,],\n",
+      "[0.009602402336895466,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20496279001235962,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5739211440086365,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6402368545532227,],\n",
+      "[-0.12667332589626312,],\n",
+      "[0.5524845719337463,],\n",
+      "[-1.0005953311920166,],\n",
+      "[0.03485004976391792,],\n",
+      "[0.45739632844924927,],\n",
+      "[-0.30948424339294434,],\n",
+      "[1.0020935535430908,],\n",
+      "[0.8306312561035156,],\n",
+      "[-0.9396960139274597,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20553670823574066,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.023555202409625053,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6402368545532227,],\n",
+      "[-0.12667332589626312,],\n",
+      "[0.5524845719337463,],\n",
+      "[-1.0005953311920166,],\n",
+      "[0.03485004976391792,],\n",
+      "[0.45739632844924927,],\n",
+      "[-0.30948424339294434,],\n",
+      "[1.0020935535430908,],\n",
+      "[0.8306312561035156,],\n",
+      "[-0.9396960139274597,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.025189856067299843,],\n",
+      "[0.008409738540649414,],\n",
+      "[0.023214517161250114,],\n",
+      "[-0.003542582970112562,],\n",
+      "[-0.489015132188797,],\n",
+      "[-0.01858048141002655,],\n",
+      "[-0.7890096306800842,],\n",
+      "[-0.25669658184051514,],\n",
+      "[-0.06079903990030289,],\n",
+      "[0.009528503753244877,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20553670823574066,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3069540560245514,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6402620673179626,],\n",
+      "[-0.12668173015117645,],\n",
+      "[0.5524613857269287,],\n",
+      "[-1.00059175491333,],\n",
+      "[0.03533906489610672,],\n",
+      "[0.4574148952960968,],\n",
+      "[-0.3086952269077301,],\n",
+      "[1.0023502111434937,],\n",
+      "[0.8306920528411865,],\n",
+      "[-0.9397055506706238,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20584365725517273,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.030719459056854248,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6402620673179626,],\n",
+      "[-0.12668173015117645,],\n",
+      "[0.5524613857269287,],\n",
+      "[-1.00059175491333,],\n",
+      "[0.03533906489610672,],\n",
+      "[0.4574148952960968,],\n",
+      "[-0.3086952269077301,],\n",
+      "[1.0023502111434937,],\n",
+      "[0.8306920528411865,],\n",
+      "[-0.9397055506706238,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.022293442860245705,],\n",
+      "[-0.007540769409388304,],\n",
+      "[-0.021734846755862236,],\n",
+      "[0.00338706374168396,],\n",
+      "[-0.4217270612716675,],\n",
+      "[-0.01588539592921734,],\n",
+      "[-0.6688854694366455,],\n",
+      "[-0.2504718005657196,],\n",
+      "[-0.05755285173654556,],\n",
+      "[0.00945803988724947,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20584365725517273,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3505389392375946,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6402397751808167,],\n",
+      "[-0.12667419016361237,],\n",
+      "[0.5524831414222717,],\n",
+      "[-1.0005950927734375,],\n",
+      "[0.0357607901096344,],\n",
+      "[0.45743077993392944,],\n",
+      "[-0.30802634358406067,],\n",
+      "[1.0026006698608398,],\n",
+      "[0.8307496309280396,],\n",
+      "[-0.9397150278091431,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20549312233924866,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.21765762567520142,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6402397751808167,],\n",
+      "[-0.12667419016361237,],\n",
+      "[0.5524831414222717,],\n",
+      "[-1.0005950927734375,],\n",
+      "[0.0357607901096344,],\n",
+      "[0.45743077993392944,],\n",
+      "[-0.30802634358406067,],\n",
+      "[1.0026006698608398,],\n",
+      "[0.8307496309280396,],\n",
+      "[-0.9397150278091431,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0453178845345974,],\n",
+      "[-0.01563439704477787,],\n",
+      "[-0.04716736078262329,],\n",
+      "[0.007518985774368048,],\n",
+      "[-0.3649892807006836,],\n",
+      "[-0.013242965564131737,],\n",
+      "[-0.5620121955871582,],\n",
+      "[-0.24652427434921265,],\n",
+      "[-0.05598090589046478,],\n",
+      "[0.00943666510283947,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20549312233924866,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9330750703811646,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401944756507874,],\n",
+      "[-0.12665855884552002,],\n",
+      "[0.5525302886962891,],\n",
+      "[-1.0006026029586792,],\n",
+      "[0.036125779151916504,],\n",
+      "[0.4574440121650696,],\n",
+      "[-0.3074643313884735,],\n",
+      "[1.0028471946716309,],\n",
+      "[0.8308055996894836,],\n",
+      "[-0.9397244453430176,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2045600414276123,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.26708075404167175,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401944756507874,],\n",
+      "[-0.12665855884552002,],\n",
+      "[0.5525302886962891,],\n",
+      "[-1.0006026029586792,],\n",
+      "[0.036125779151916504,],\n",
+      "[0.4574440121650696,],\n",
+      "[-0.3074643313884735,],\n",
+      "[1.0028471946716309,],\n",
+      "[0.8308055996894836,],\n",
+      "[-0.9397244453430176,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.03790593892335892,],\n",
+      "[-0.013408319093286991,],\n",
+      "[-0.04240047559142113,],\n",
+      "[0.006923920474946499,],\n",
+      "[-0.3577357828617096,],\n",
+      "[-0.012514432892203331,],\n",
+      "[-0.5431985259056091,],\n",
+      "[-0.24635851383209229,],\n",
+      "[-0.05646311119198799,],\n",
+      "[0.00945513416081667,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2045600414276123,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0335978269577026,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401565670967102,],\n",
+      "[-0.12664514780044556,],\n",
+      "[0.5525726675987244,],\n",
+      "[-1.0006095170974731,],\n",
+      "[0.0364835150539875,],\n",
+      "[0.4574565291404724,],\n",
+      "[-0.306921124458313,],\n",
+      "[1.0030936002731323,],\n",
+      "[0.8308620452880859,],\n",
+      "[-0.9397339224815369,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20352643728256226,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07521006464958191,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401565670967102,],\n",
+      "[-0.12664514780044556,],\n",
+      "[0.5525726675987244,],\n",
+      "[-1.0006095170974731,],\n",
+      "[0.0364835150539875,],\n",
+      "[0.4574565291404724,],\n",
+      "[-0.306921124458313,],\n",
+      "[1.0030936002731323,],\n",
+      "[0.8308620452880859,],\n",
+      "[-0.9397339224815369,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.015065331012010574,],\n",
+      "[-0.005478707142174244,],\n",
+      "[-0.018191775307059288,],\n",
+      "[0.003046711441129446,],\n",
+      "[-0.4056192636489868,],\n",
+      "[-0.01414540782570839,],\n",
+      "[-0.6315473318099976,],\n",
+      "[-0.2488010972738266,],\n",
+      "[-0.05796115845441818,],\n",
+      "[0.009485555812716484,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20352643728256226,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.548488438129425,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.640141487121582,],\n",
+      "[-0.12663966417312622,],\n",
+      "[0.5525908470153809,],\n",
+      "[-1.0006126165390015,],\n",
+      "[0.03688913583755493,],\n",
+      "[0.45747068524360657,],\n",
+      "[-0.30628958344459534,],\n",
+      "[1.0033423900604248,],\n",
+      "[0.8309199810028076,],\n",
+      "[-0.9397433996200562,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2029779553413391,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.017078371718525887,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.640141487121582,],\n",
+      "[-0.12663966417312622,],\n",
+      "[0.5525908470153809,],\n",
+      "[-1.0006126165390015,],\n",
+      "[0.03688913583755493,],\n",
+      "[0.45747068524360657,],\n",
+      "[-0.30628958344459534,],\n",
+      "[1.0033423900604248,],\n",
+      "[0.8309199810028076,],\n",
+      "[-0.9397433996200562,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005346586927771568,],\n",
+      "[0.0019992010202258825,],\n",
+      "[0.006988563574850559,],\n",
+      "[-0.0012015573447570205,],\n",
+      "[-0.47985896468162537,],\n",
+      "[-0.017135510221123695,],\n",
+      "[-0.779439389705658,],\n",
+      "[-0.2517968416213989,],\n",
+      "[-0.05928632989525795,],\n",
+      "[0.009507069364190102,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2029779553413391,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2613682150840759,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401468515396118,],\n",
+      "[-0.1266416609287262,],\n",
+      "[0.5525838732719421,],\n",
+      "[-1.000611424446106,],\n",
+      "[0.03736899420619011,],\n",
+      "[0.4574878215789795,],\n",
+      "[-0.30551013350486755,],\n",
+      "[1.003594160079956,],\n",
+      "[0.8309792876243591,],\n",
+      "[-0.9397529363632202,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2032393217086792,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2164885699748993,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401468515396118,],\n",
+      "[-0.1266416609287262,],\n",
+      "[0.5525838732719421,],\n",
+      "[-1.000611424446106,],\n",
+      "[0.03736899420619011,],\n",
+      "[0.4574878215789795,],\n",
+      "[-0.30551013350486755,],\n",
+      "[1.003594160079956,],\n",
+      "[0.8309792876243591,],\n",
+      "[-0.9397529363632202,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.01412166841328144,],\n",
+      "[0.005421786103397608,],\n",
+      "[0.020012613385915756,],\n",
+      "[-0.00353518920019269,],\n",
+      "[-0.5371765494346619,],\n",
+      "[-0.019861003383994102,],\n",
+      "[-0.9019452333450317,],\n",
+      "[-0.2538662254810333,],\n",
+      "[-0.05984930321574211,],\n",
+      "[0.009513629600405693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2032393217086792,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9305660128593445,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401609778404236,],\n",
+      "[-0.12664708495140076,],\n",
+      "[0.5525638461112976,],\n",
+      "[-1.0006078481674194,],\n",
+      "[0.03790616989135742,],\n",
+      "[0.4575076699256897,],\n",
+      "[-0.30460819602012634,],\n",
+      "[1.0038480758666992,],\n",
+      "[0.8310391306877136,],\n",
+      "[-0.9397624731063843,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2041698843240738,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2879065275192261,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401609778404236,],\n",
+      "[-0.12664708495140076,],\n",
+      "[0.5525638461112976,],\n",
+      "[-1.0006078481674194,],\n",
+      "[0.03790616989135742,],\n",
+      "[0.4575076699256897,],\n",
+      "[-0.30460819602012634,],\n",
+      "[1.0038480758666992,],\n",
+      "[0.8310391306877136,],\n",
+      "[-0.9397624731063843,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.012045510113239288,],\n",
+      "[0.004742508754134178,],\n",
+      "[0.018526805564761162,],\n",
+      "[-0.0033646642696112394,],\n",
+      "[-0.546591579914093,],\n",
+      "[-0.020847642794251442,],\n",
+      "[-0.9284606575965881,],\n",
+      "[-0.25458288192749023,],\n",
+      "[-0.059720005840063095,],\n",
+      "[0.009509850293397903,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2041698843240738,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0731383562088013,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401730179786682,],\n",
+      "[-0.1266518235206604,],\n",
+      "[0.5525453090667725,],\n",
+      "[-1.000604510307312,],\n",
+      "[0.038452763110399246,],\n",
+      "[0.4575285315513611,],\n",
+      "[-0.3036797344684601,],\n",
+      "[1.0041027069091797,],\n",
+      "[0.8310988545417786,],\n",
+      "[-0.9397720098495483,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20524302124977112,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09552823007106781,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401730179786682,],\n",
+      "[-0.1266518235206604,],\n",
+      "[0.5525453090667725,],\n",
+      "[-1.000604510307312,],\n",
+      "[0.038452763110399246,],\n",
+      "[0.4575285315513611,],\n",
+      "[-0.3036797344684601,],\n",
+      "[1.0041027069091797,],\n",
+      "[0.8310988545417786,],\n",
+      "[-0.9397720098495483,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005119509994983673,],\n",
+      "[0.0020672460086643696,],\n",
+      "[0.008553721942007542,],\n",
+      "[-0.0015977186849340796,],\n",
+      "[-0.5069056749343872,],\n",
+      "[-0.019498590379953384,],\n",
+      "[-0.8457538485527039,],\n",
+      "[-0.25425687432289124,],\n",
+      "[-0.059291236102581024,],\n",
+      "[0.00950307585299015,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20524302124977112,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6181533336639404,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401781439781189,],\n",
+      "[-0.12665389478206635,],\n",
+      "[0.5525367259979248,],\n",
+      "[-1.0006029605865479,],\n",
+      "[0.03895966708660126,],\n",
+      "[0.45754802227020264,],\n",
+      "[-0.3028339743614197,],\n",
+      "[1.0043569803237915,],\n",
+      "[0.8311581611633301,],\n",
+      "[-0.9397814869880676,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2058611810207367,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.00533501198515296,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401781439781189,],\n",
+      "[-0.12665389478206635,],\n",
+      "[0.5525367259979248,],\n",
+      "[-1.0006029605865479,],\n",
+      "[0.03895966708660126,],\n",
+      "[0.45754802227020264,],\n",
+      "[-0.3028339743614197,],\n",
+      "[1.0043569803237915,],\n",
+      "[0.8311581611633301,],\n",
+      "[-0.9397814869880676,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0008908259333111346,],\n",
+      "[-0.00036956529947929084,],\n",
+      "[-0.0016182982362806797,],\n",
+      "[0.00031092524295672774,],\n",
+      "[-0.44575005769729614,],\n",
+      "[-0.016669083386659622,],\n",
+      "[-0.7064606547355652,],\n",
+      "[-0.25344571471214294,],\n",
+      "[-0.058924149721860886,],\n",
+      "[0.00949830375611782,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2058611810207367,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.14608220756053925,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401772499084473,],\n",
+      "[-0.1266535222530365,],\n",
+      "[0.5525383353233337,],\n",
+      "[-1.0006033182144165,],\n",
+      "[0.03940541669726372,],\n",
+      "[0.45756468176841736,],\n",
+      "[-0.30212751030921936,],\n",
+      "[1.0046104192733765,],\n",
+      "[0.8312171101570129,],\n",
+      "[-0.9397909641265869,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2057151049375534,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13921163976192474,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401772499084473,],\n",
+      "[-0.1266535222530365,],\n",
+      "[0.5525383353233337,],\n",
+      "[-1.0006033182144165,],\n",
+      "[0.03940541669726372,],\n",
+      "[0.45756468176841736,],\n",
+      "[-0.30212751030921936,],\n",
+      "[1.0046104192733765,],\n",
+      "[0.8312171101570129,],\n",
+      "[-0.9397909641265869,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0033451239578425884,],\n",
+      "[-0.0014293133281171322,],\n",
+      "[-0.006612624041736126,],\n",
+      "[0.0013067005202174187,],\n",
+      "[-0.4014916718006134,],\n",
+      "[-0.014265039004385471,],\n",
+      "[-0.5971036553382874,],\n",
+      "[-0.252603143453598,],\n",
+      "[-0.058776069432497025,],\n",
+      "[0.009496877901256084,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2057151049375534,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7462207674980164,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401739120483398,],\n",
+      "[-0.1266520917415619,],\n",
+      "[0.5525449514389038,],\n",
+      "[-1.0006046295166016,],\n",
+      "[0.03980690985918045,],\n",
+      "[0.45757895708084106,],\n",
+      "[-0.3015304207801819,],\n",
+      "[1.0048630237579346,],\n",
+      "[0.8312758803367615,],\n",
+      "[-0.9398004412651062,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2049688845872879,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.16431336104869843,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401739120483398,],\n",
+      "[-0.1266520917415619,],\n",
+      "[0.5525449514389038,],\n",
+      "[-1.0006046295166016,],\n",
+      "[0.03980690985918045,],\n",
+      "[0.45757895708084106,],\n",
+      "[-0.3015304207801819,],\n",
+      "[1.0048630237579346,],\n",
+      "[0.8312758803367615,],\n",
+      "[-0.9398004412651062,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0026684636250138283,],\n",
+      "[-0.0011765516828745604,],\n",
+      "[-0.0057432809844613075,],\n",
+      "[0.0011670857202261686,],\n",
+      "[-0.3992704451084137,],\n",
+      "[-0.013848943635821342,],\n",
+      "[-0.5852338075637817,],\n",
+      "[-0.2519857585430145,],\n",
+      "[-0.0588158518075943,],\n",
+      "[0.009497661143541336,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2049688845872879,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8107122182846069,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.640171229839325,],\n",
+      "[-0.12665091454982758,],\n",
+      "[0.5525506734848022,],\n",
+      "[-1.000605821609497,],\n",
+      "[0.040206179022789,],\n",
+      "[0.45759281516075134,],\n",
+      "[-0.3009451925754547,],\n",
+      "[1.005115032196045,],\n",
+      "[0.8313347101211548,],\n",
+      "[-0.9398099184036255,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20415817201137543,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.021764032542705536,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.640171229839325,],\n",
+      "[-0.12665091454982758,],\n",
+      "[0.5525506734848022,],\n",
+      "[-1.000605821609497,],\n",
+      "[0.040206179022789,],\n",
+      "[0.45759281516075134,],\n",
+      "[-0.3009451925754547,],\n",
+      "[1.005115032196045,],\n",
+      "[0.8313347101211548,],\n",
+      "[-0.9398099184036255,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0007125834817998111,],\n",
+      "[-0.00032428649137727916,],\n",
+      "[-0.0016700560227036476,],\n",
+      "[0.0003490022791083902,],\n",
+      "[-0.43677830696105957,],\n",
+      "[-0.01557473186403513,],\n",
+      "[-0.6786532998085022,],\n",
+      "[-0.25174447894096375,],\n",
+      "[-0.05892976000905037,],\n",
+      "[0.009499039500951767,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20415817201137543,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.29505226016044617,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401705145835876,],\n",
+      "[-0.1266505867242813,],\n",
+      "[0.552552342414856,],\n",
+      "[-1.0006061792373657,],\n",
+      "[0.040642958134412766,],\n",
+      "[0.4576084017753601,],\n",
+      "[-0.3002665340900421,],\n",
+      "[1.0053668022155762,],\n",
+      "[0.8313936591148376,],\n",
+      "[-0.9398193955421448,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20386311411857605,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.060366272926330566,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401705145835876,],\n",
+      "[-0.1266505867242813,],\n",
+      "[0.552552342414856,],\n",
+      "[-1.0006061792373657,],\n",
+      "[0.040642958134412766,],\n",
+      "[0.4576084017753601,],\n",
+      "[-0.3002665340900421,],\n",
+      "[1.0053668022155762,],\n",
+      "[0.8313936591148376,],\n",
+      "[-0.9398193955421448,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0008704139618203044,],\n",
+      "[0.00040826265467330813,],\n",
+      "[0.002220723545178771,],\n",
+      "[-0.000477404537377879,],\n",
+      "[-0.4884140193462372,],\n",
+      "[-0.018324514850974083,],\n",
+      "[-0.8210306763648987,],\n",
+      "[-0.25198978185653687,],\n",
+      "[-0.059020694345235825,],\n",
+      "[0.009499919600784779,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20386311411857605,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4913903772830963,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401714086532593,],\n",
+      "[-0.12665098905563354,],\n",
+      "[0.5525501370429993,],\n",
+      "[-1.0006057024002075,],\n",
+      "[0.04113137349486351,],\n",
+      "[0.45762673020362854,],\n",
+      "[-0.2994455099105835,],\n",
+      "[1.0056188106536865,],\n",
+      "[0.8314526677131653,],\n",
+      "[-0.9398288726806641,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20435450971126556,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2880457639694214,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401714086532593,],\n",
+      "[-0.12665098905563354,],\n",
+      "[0.5525501370429993,],\n",
+      "[-1.0006057024002075,],\n",
+      "[0.04113137349486351,],\n",
+      "[0.45762673020362854,],\n",
+      "[-0.2994455099105835,],\n",
+      "[1.0056188106536865,],\n",
+      "[0.8314526677131653,],\n",
+      "[-0.9398288726806641,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0013941394863650203,],\n",
+      "[0.0006724906852468848,],\n",
+      "[0.0038701791781932116,],\n",
+      "[-0.0008563029696233571,],\n",
+      "[-0.5228577852249146,],\n",
+      "[-0.020522234961390495,],\n",
+      "[-0.9261103272438049,],\n",
+      "[-0.25270941853523254,],\n",
+      "[-0.05905060097575188,],\n",
+      "[0.009500058367848396,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20435450971126556,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0733965635299683,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401727795600891,],\n",
+      "[-0.12665165960788727,],\n",
+      "[0.5525462627410889,],\n",
+      "[-1.0006048679351807,],\n",
+      "[0.041654232889413834,],\n",
+      "[0.45764726400375366,],\n",
+      "[-0.29851940274238586,],\n",
+      "[1.0058715343475342,],\n",
+      "[0.8315117359161377,],\n",
+      "[-0.9398383498191833,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20542789995670319,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.30177736282348633,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401727795600891,],\n",
+      "[-0.12665165960788727,],\n",
+      "[0.5525462627410889,],\n",
+      "[-1.0006048679351807,],\n",
+      "[0.041654232889413834,],\n",
+      "[0.45764726400375366,],\n",
+      "[-0.29851940274238586,],\n",
+      "[1.0058715343475342,],\n",
+      "[0.8315117359161377,],\n",
+      "[-0.9398383498191833,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.001046080724336207,],\n",
+      "[0.0005180948064662516,],\n",
+      "[0.0031583947129547596,],\n",
+      "[-0.0007195108337327838,],\n",
+      "[-0.5215479135513306,],\n",
+      "[-0.02092817984521389,],\n",
+      "[-0.9303877949714661,],\n",
+      "[-0.25368532538414,],\n",
+      "[-0.05903145298361778,],\n",
+      "[0.009499800391495228,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20542789995670319,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.09868586063385,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401738524436951,],\n",
+      "[-0.12665218114852905,],\n",
+      "[0.5525431036949158,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.0421757809817791,],\n",
+      "[0.4576681852340698,],\n",
+      "[-0.2975890040397644,],\n",
+      "[1.0061252117156982,],\n",
+      "[0.8315707445144653,],\n",
+      "[-0.9398478269577026,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20652659237384796,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07622041553258896,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401738524436951,],\n",
+      "[-0.12665218114852905,],\n",
+      "[0.5525431036949158,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.0421757809817791,],\n",
+      "[0.4576681852340698,],\n",
+      "[-0.2975890040397644,],\n",
+      "[1.0061252117156982,],\n",
+      "[0.8315707445144653,],\n",
+      "[-0.9398478269577026,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00038530558231286705,],\n",
+      "[0.00019592219905462116,],\n",
+      "[0.0012651775032281876,],\n",
+      "[-0.0002967689069919288,],\n",
+      "[-0.48810938000679016,],\n",
+      "[-0.019309978932142258,],\n",
+      "[-0.8322182297706604,],\n",
+      "[-0.2545170187950134,],\n",
+      "[-0.058995191007852554,],\n",
+      "[0.009499470703303814,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20652659237384796,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5521602034568787,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401742100715637,],\n",
+      "[-0.12665237486362457,],\n",
+      "[0.5525418519973755,],\n",
+      "[-1.0006039142608643,],\n",
+      "[0.04266389086842537,],\n",
+      "[0.45768749713897705,],\n",
+      "[-0.296756774187088,],\n",
+      "[1.0063797235488892,],\n",
+      "[0.831629753112793,],\n",
+      "[-0.9398573040962219,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.207078754901886,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.013762828893959522,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401742100715637,],\n",
+      "[-0.12665237486362457,],\n",
+      "[0.5525418519973755,],\n",
+      "[-1.0006039142608643,],\n",
+      "[0.04266389086842537,],\n",
+      "[0.45768749713897705,],\n",
+      "[-0.296756774187088,],\n",
+      "[1.0063797235488892,],\n",
+      "[0.831629753112793,],\n",
+      "[-0.9398573040962219,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00011997184628853574,],\n",
+      "[-6.272719474509358e-05,],\n",
+      "[-0.0004285539616830647,],\n",
+      "[0.00010347087663831189,],\n",
+      "[-0.4446919858455658,],\n",
+      "[-0.016711778938770294,],\n",
+      "[-0.6921574473381042,],\n",
+      "[-0.25483304262161255,],\n",
+      "[-0.05896798148751259,],\n",
+      "[0.009499173611402512,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.207078754901886,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.23463033139705658,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401740908622742,],\n",
+      "[-0.1266523152589798,],\n",
+      "[0.5525422692298889,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.043108582496643066,],\n",
+      "[0.45770421624183655,],\n",
+      "[-0.2960646152496338,],\n",
+      "[1.0066345930099487,],\n",
+      "[0.8316887021064758,],\n",
+      "[-0.9398667812347412,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20684412121772766,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1539180427789688,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401740908622742,],\n",
+      "[-0.1266523152589798,],\n",
+      "[0.5525422692298889,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.043108582496643066,],\n",
+      "[0.45770421624183655,],\n",
+      "[-0.2960646152496338,],\n",
+      "[1.0066345930099487,],\n",
+      "[0.8316887021064758,],\n",
+      "[-0.9398667812347412,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00029393701697699726,],\n",
+      "[-0.00015837338287383318,],\n",
+      "[-0.0011428208090364933,],\n",
+      "[0.0002838555083144456,],\n",
+      "[-0.4175432026386261,],\n",
+      "[-0.01482409331947565,],\n",
+      "[-0.5948280692100525,],\n",
+      "[-0.25449731945991516,],\n",
+      "[-0.05895855650305748,],\n",
+      "[0.009499178268015385,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20684412121772766,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7846474051475525,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737928390503,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525434017181396,],\n",
+      "[-1.000604271888733,],\n",
+      "[0.04352612420916557,],\n",
+      "[0.45771902799606323,],\n",
+      "[-0.2954697906970978,],\n",
+      "[1.0068891048431396,],\n",
+      "[0.8317476511001587,],\n",
+      "[-0.9398762583732605,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20605947077274323,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14629821479320526,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737928390503,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525434017181396,],\n",
+      "[-1.000604271888733,],\n",
+      "[0.04352612420916557,],\n",
+      "[0.45771902799606323,],\n",
+      "[-0.2954697906970978,],\n",
+      "[1.0068891048431396,],\n",
+      "[0.8317476511001587,],\n",
+      "[-0.9398762583732605,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00020992632198613137,],\n",
+      "[-0.00011673692642943934,],\n",
+      "[-0.0008886725991033018,],\n",
+      "[0.00022698086104355752,],\n",
+      "[-0.42064473032951355,],\n",
+      "[-0.014784913510084152,],\n",
+      "[-0.5981531739234924,],\n",
+      "[-0.2537490725517273,],\n",
+      "[-0.05896276235580444,],\n",
+      "[0.00949918944388628,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20605947077274323,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.764978289604187,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401735544204712,],\n",
+      "[-0.12665203213691711,],\n",
+      "[0.5525442957878113,],\n",
+      "[-1.000604510307312,],\n",
+      "[0.0439467690885067,],\n",
+      "[0.45773380994796753,],\n",
+      "[-0.29487162828445435,],\n",
+      "[1.0071429014205933,],\n",
+      "[0.8318066000938416,],\n",
+      "[-0.9398857355117798,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20529448986053467,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.00893666036427021,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401735544204712,],\n",
+      "[-0.12665203213691711,],\n",
+      "[0.5525442957878113,],\n",
+      "[-1.000604510307312,],\n",
+      "[0.0439467690885067,],\n",
+      "[0.45773380994796753,],\n",
+      "[-0.29487162828445435,],\n",
+      "[1.0071429014205933,],\n",
+      "[0.8318066000938416,],\n",
+      "[-0.9398857355117798,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.8005851820344105e-05,],\n",
+      "[-2.1810943508171476e-05,],\n",
+      "[-0.00017516819934826344,],\n",
+      "[4.600758620654233e-05,],\n",
+      "[-0.4485089182853699,],\n",
+      "[-0.01647854410111904,],\n",
+      "[-0.6994783878326416,],\n",
+      "[-0.2530859112739563,],\n",
+      "[-0.058971889317035675,],\n",
+      "[0.009499290026724339,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20529448986053467,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.18906782567501068,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401734948158264,],\n",
+      "[-0.12665201723575592,],\n",
+      "[0.5525444746017456,],\n",
+      "[-1.000604510307312,],\n",
+      "[0.04439527913928032,],\n",
+      "[0.4577502906322479,],\n",
+      "[-0.29417213797569275,],\n",
+      "[1.0073959827423096,],\n",
+      "[0.8318655490875244,],\n",
+      "[-0.9398952126502991,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20510542392730713,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08740561455488205,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401734948158264,],\n",
+      "[-0.12665201723575592,],\n",
+      "[0.5525444746017456,],\n",
+      "[-1.000604510307312,],\n",
+      "[0.04439527913928032,],\n",
+      "[0.4577502906322479,],\n",
+      "[-0.29417213797569275,],\n",
+      "[1.0073959827423096,],\n",
+      "[0.8318655490875244,],\n",
+      "[-0.9398952126502991,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.706362132215872e-05,],\n",
+      "[5.1481030823197216e-05,],\n",
+      "[0.000436687667388469,],\n",
+      "[-0.0001179921964649111,],\n",
+      "[-0.481985867023468,],\n",
+      "[-0.01879759505391121,],\n",
+      "[-0.8364680409431458,],\n",
+      "[-0.25298038125038147,],\n",
+      "[-0.058978188782930374,],\n",
+      "[0.009499290958046913,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20510542392730713,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5912893414497375,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401735544204712,],\n",
+      "[-0.1266520619392395,],\n",
+      "[0.5525440573692322,],\n",
+      "[-1.0006043910980225,],\n",
+      "[0.04487726464867592,],\n",
+      "[0.45776909589767456,],\n",
+      "[-0.2933356761932373,],\n",
+      "[1.0076489448547363,],\n",
+      "[0.8319244980812073,],\n",
+      "[-0.9399046897888184,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2056967169046402,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.30226173996925354,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401735544204712,],\n",
+      "[-0.1266520619392395,],\n",
+      "[0.5525440573692322,],\n",
+      "[-1.0006043910980225,],\n",
+      "[0.04487726464867592,],\n",
+      "[0.45776909589767456,],\n",
+      "[-0.2933356761932373,],\n",
+      "[1.0076489448547363,],\n",
+      "[0.8319244980812073,],\n",
+      "[-0.9399046897888184,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00011859287042170763,],\n",
+      "[7.209648902062327e-05,],\n",
+      "[0.0006469158106483519,],\n",
+      "[-0.000179928116267547,],\n",
+      "[-0.5010488629341125,],\n",
+      "[-0.020387036725878716,],\n",
+      "[-0.9244846701622009,],\n",
+      "[-0.25360122323036194,],\n",
+      "[-0.05897975340485573,],\n",
+      "[0.00949928816407919,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2056967169046402,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0995670557022095,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401736736297607,],\n",
+      "[-0.12665213644504547,],\n",
+      "[0.5525434017181396,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04537831246852875,],\n",
+      "[0.45778948068618774,],\n",
+      "[-0.2924111783504486,],\n",
+      "[1.0079025030136108,],\n",
+      "[0.8319835066795349,],\n",
+      "[-0.9399141669273376,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2067962884902954,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2621975243091583,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401736736297607,],\n",
+      "[-0.12665213644504547,],\n",
+      "[0.5525434017181396,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04537831246852875,],\n",
+      "[0.45778948068618774,],\n",
+      "[-0.2924111783504486,],\n",
+      "[1.0079025030136108,],\n",
+      "[0.8319835066795349,],\n",
+      "[-0.9399141669273376,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.090433402685449e-05,],\n",
+      "[5.049626270192675e-05,],\n",
+      "[0.0004797680885531008,],\n",
+      "[-0.00013741134898737073,],\n",
+      "[-0.49600568413734436,],\n",
+      "[-0.020354345440864563,],\n",
+      "[-0.9098567962646484,],\n",
+      "[-0.2546795606613159,],\n",
+      "[-0.058977991342544556,],\n",
+      "[0.00949928443878889,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2067962884902954,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0241042375564575,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665218114852905,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.04587431997060776,],\n",
+      "[0.45780983567237854,],\n",
+      "[-0.29150131344795227,],\n",
+      "[1.0081571340560913,],\n",
+      "[0.8320424556732178,],\n",
+      "[-0.9399236440658569,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20782038569450378,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04151994735002518,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665218114852905,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.04587431997060776,],\n",
+      "[0.45780983567237854,],\n",
+      "[-0.29150131344795227,],\n",
+      "[1.0081571340560913,],\n",
+      "[0.8320424556732178,],\n",
+      "[-0.9399236440658569,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.3581324057886377e-05,],\n",
+      "[1.5113795598153956e-05,],\n",
+      "[0.00015202254871837795,],\n",
+      "[-4.483469820115715e-05,],\n",
+      "[-0.4718818962574005,],\n",
+      "[-0.018765589222311974,],\n",
+      "[-0.8034884333610535,],\n",
+      "[-0.2556557059288025,],\n",
+      "[-0.05897507816553116,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20782038569450378,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.40752890706062317,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665219604969025,],\n",
+      "[0.5525427460670471,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.04634620249271393,],\n",
+      "[0.4578286111354828,],\n",
+      "[-0.2906978130340576,],\n",
+      "[1.0084128379821777,],\n",
+      "[0.8321014046669006,],\n",
+      "[-0.9399331212043762,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20822791755199432,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0359354168176651,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665219604969025,],\n",
+      "[0.5525427460670471,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.04634620249271393,],\n",
+      "[0.4578286111354828,],\n",
+      "[-0.2906978130340576,],\n",
+      "[1.0084128379821777,],\n",
+      "[0.8321014046669006,],\n",
+      "[-0.9399331212043762,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.6068401237134822e-05,],\n",
+      "[-1.0593430488370359e-05,],\n",
+      "[-0.00011266675574006513,],\n",
+      "[3.419785207370296e-05,],\n",
+      "[-0.4444064199924469,],\n",
+      "[-0.016651874408125877,],\n",
+      "[-0.6709849834442139,],\n",
+      "[-0.2560274004936218,],\n",
+      "[-0.05897323414683342,],\n",
+      "[0.009499281644821167,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20822791755199432,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3791331648826599,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665218114852905,],\n",
+      "[0.5525428652763367,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.046790607273578644,],\n",
+      "[0.4578452706336975,],\n",
+      "[-0.29002681374549866,],\n",
+      "[1.0086688995361328,],\n",
+      "[0.8321603536605835,],\n",
+      "[-0.9399425983428955,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20784878730773926,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.18560022115707397,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665218114852905,],\n",
+      "[0.5525428652763367,],\n",
+      "[-1.0006040334701538,],\n",
+      "[0.046790607273578644,],\n",
+      "[0.4578452706336975,],\n",
+      "[-0.29002681374549866,],\n",
+      "[1.0086688995361328,],\n",
+      "[0.8321603536605835,],\n",
+      "[-0.9399425983428955,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.674640745681245e-05,],\n",
+      "[-1.8177541278419085e-05,],\n",
+      "[-0.00020409785793162882,],\n",
+      "[6.371727795340121e-05,],\n",
+      "[-0.42980247735977173,],\n",
+      "[-0.015343071892857552,],\n",
+      "[-0.5904523134231567,],\n",
+      "[-0.2556334137916565,],\n",
+      "[-0.05897285044193268,],\n",
+      "[0.009499281644821167,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20784878730773926,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8616268038749695,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665216624736786,],\n",
+      "[0.552543044090271,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04722040891647339,],\n",
+      "[0.4578606188297272,],\n",
+      "[-0.28943637013435364,],\n",
+      "[1.0089244842529297,],\n",
+      "[0.8322193026542664,],\n",
+      "[-0.9399520754814148,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2069871574640274,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1407288908958435,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665216624736786,],\n",
+      "[0.552543044090271,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04722040891647339,],\n",
+      "[0.4578606188297272,],\n",
+      "[-0.28943637013435364,],\n",
+      "[1.0089244842529297,],\n",
+      "[0.8322193026542664,],\n",
+      "[-0.9399520754814148,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.7058002413250506e-05,],\n",
+      "[-1.196697212435538e-05,],\n",
+      "[-0.00014171525253914297,],\n",
+      "[4.548564902506769e-05,],\n",
+      "[-0.4348278343677521,],\n",
+      "[-0.01556631550192833,],\n",
+      "[-0.6084946990013123,],\n",
+      "[-0.2547808289527893,],\n",
+      "[-0.058973200619220734,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2069871574640274,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7502774596214294,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525431632995605,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04765523597598076,],\n",
+      "[0.45787617564201355,],\n",
+      "[-0.2888278663158417,],\n",
+      "[1.0091792345046997,],\n",
+      "[0.8322782516479492,],\n",
+      "[-0.9399615526199341,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20623688399791718,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0033715874888002872,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525431632995605,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04765523597598076,],\n",
+      "[0.45787617564201355,],\n",
+      "[-0.2888278663158417,],\n",
+      "[1.0091792345046997,],\n",
+      "[0.8322782516479492,],\n",
+      "[-0.9399615526199341,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.9338046968186973e-06,],\n",
+      "[-1.3999660950503312e-06,],\n",
+      "[-1.7489322999608703e-05,],\n",
+      "[5.771740688942373e-06,],\n",
+      "[-0.4538845419883728,],\n",
+      "[-0.017050238326191902,],\n",
+      "[-0.7139307260513306,],\n",
+      "[-0.25404852628707886,],\n",
+      "[-0.05897393077611923,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20623688399791718,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.11613070964813232,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525431632995605,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04810912162065506,],\n",
+      "[0.4578932225704193,],\n",
+      "[-0.28811392188072205,],\n",
+      "[1.0094332695007324,],\n",
+      "[0.8323372006416321,],\n",
+      "[-0.9399710297584534,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20612075924873352,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10621268302202225,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525431632995605,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04810912162065506,],\n",
+      "[0.4578932225704193,],\n",
+      "[-0.28811392188072205,],\n",
+      "[1.0094332695007324,],\n",
+      "[0.8323372006416321,],\n",
+      "[-0.9399710297584534,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.949570317578036e-06,],\n",
+      "[5.928355676587671e-06,],\n",
+      "[7.822483166819438e-05,],\n",
+      "[-2.6557067030807957e-05,],\n",
+      "[-0.47413063049316406,],\n",
+      "[-0.01882803998887539,],\n",
+      "[-0.8413675427436829,],\n",
+      "[-0.25394731760025024,],\n",
+      "[-0.05897430330514908,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20612075924873352,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6518064141273499,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525431036949158,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.0485832504928112,],\n",
+      "[0.45791205763816833,],\n",
+      "[-0.28727254271507263,],\n",
+      "[1.0096871852874756,],\n",
+      "[0.8323961496353149,],\n",
+      "[-0.9399805068969727,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2067725658416748,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2932267487049103,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525431036949158,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.0485832504928112,],\n",
+      "[0.45791205763816833,],\n",
+      "[-0.28727254271507263,],\n",
+      "[1.0096871852874756,],\n",
+      "[0.8323961496353149,],\n",
+      "[-0.9399805068969727,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-9.674211469246075e-06,],\n",
+      "[7.416008884320036e-06,],\n",
+      "[0.00010350943921366706,],\n",
+      "[-3.61737547791563e-05,],\n",
+      "[-0.4837666153907776,],\n",
+      "[-0.01985618844628334,],\n",
+      "[-0.9105300307273865,],\n",
+      "[-0.2546020746231079,],\n",
+      "[-0.058974310755729675,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2067725658416748,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.083008885383606,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04906701669096947,],\n",
+      "[0.45793190598487854,],\n",
+      "[-0.28636202216148376,],\n",
+      "[1.009941816329956,],\n",
+      "[0.8324550986289978,],\n",
+      "[-0.9399899840354919,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20785558223724365,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2077501267194748,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04906701669096947,],\n",
+      "[0.45793190598487854,],\n",
+      "[-0.28636202216148376,],\n",
+      "[1.009941816329956,],\n",
+      "[0.8324550986289978,],\n",
+      "[-0.9399899840354919,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.964070624031592e-06,],\n",
+      "[4.693948540079873e-06,],\n",
+      "[6.93600086378865e-05,],\n",
+      "[-2.4960729206213728e-05,],\n",
+      "[-0.4783387780189514,],\n",
+      "[-0.019576190039515495,],\n",
+      "[-0.8798332214355469,],\n",
+      "[-0.25567907094955444,],\n",
+      "[-0.05897431820631027,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20785558223724365,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9115921854972839,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04954535514116287,],\n",
+      "[0.45795148611068726,],\n",
+      "[-0.2854821979999542,],\n",
+      "[1.0101975202560425,],\n",
+      "[0.8325140476226807,],\n",
+      "[-0.9399994611740112,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20876717567443848,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.014109023846685886,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.04954535514116287,],\n",
+      "[0.45795148611068726,],\n",
+      "[-0.2854821979999542,],\n",
+      "[1.0101975202560425,],\n",
+      "[0.8325140476226807,],\n",
+      "[-0.9399994611740112,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.1383577884771512e-06,],\n",
+      "[9.202778414874047e-07,],\n",
+      "[1.4391484910447616e-05,],\n",
+      "[-5.332442924554925e-06,],\n",
+      "[-0.46244633197784424,],\n",
+      "[-0.01821758784353733,],\n",
+      "[-0.7720271944999695,],\n",
+      "[-0.25658467411994934,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20876717567443848,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2375631481409073,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05000780150294304,],\n",
+      "[0.45796969532966614,],\n",
+      "[-0.284710168838501,],\n",
+      "[1.0104540586471558,],\n",
+      "[0.8325729966163635,],\n",
+      "[-0.9400089383125305,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.209004744887352,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07135878503322601,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05000780150294304,],\n",
+      "[0.45796969532966614,],\n",
+      "[-0.284710168838501,],\n",
+      "[1.0104540586471558,],\n",
+      "[0.8325729966163635,],\n",
+      "[-0.9400089383125305,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.8750391745925299e-06,],\n",
+      "[-1.5599423477397067e-06,],\n",
+      "[-2.5783459932426922e-05,],\n",
+      "[9.83070185611723e-06,],\n",
+      "[-0.44633522629737854,],\n",
+      "[-0.016657480970025063,],\n",
+      "[-0.6531618237495422,],\n",
+      "[-0.2568162679672241,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.209004744887352,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5342609286308289,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05045413598418236,],\n",
+      "[0.45798635482788086,],\n",
+      "[-0.28405702114105225,],\n",
+      "[1.0107108354568481,],\n",
+      "[0.8326319456100464,],\n",
+      "[-0.9400184154510498,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20847047865390778,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22069741785526276,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429248809814,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05045413598418236,],\n",
+      "[0.45798635482788086,],\n",
+      "[-0.28405702114105225,],\n",
+      "[1.0107108354568481,],\n",
+      "[0.8326319456100464,],\n",
+      "[-0.9400184154510498,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.4151265733962646e-06,],\n",
+      "[-2.0721829514513956e-06,],\n",
+      "[-3.6145713238511235e-05,],\n",
+      "[1.417221938027069e-05,],\n",
+      "[-0.4392887055873871,],\n",
+      "[-0.01584537886083126,],\n",
+      "[-0.5908927917480469,],\n",
+      "[-0.2562790513038635,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20847047865390778,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9395701885223389,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.050893425941467285,],\n",
+      "[0.4580022096633911,],\n",
+      "[-0.2834661304950714,],\n",
+      "[1.0109671354293823,],\n",
+      "[0.8326908946037292,],\n",
+      "[-0.9400278925895691,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2075309157371521,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1354188770055771,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.050893425941467285,],\n",
+      "[0.4580022096633911,],\n",
+      "[-0.2834661304950714,],\n",
+      "[1.0109671354293823,],\n",
+      "[0.8326908946037292,],\n",
+      "[-0.9400278925895691,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.3855953966412926e-06,],\n",
+      "[-1.227442567142134e-06,],\n",
+      "[-2.2578020434593782e-05,],\n",
+      "[9.100283932639286e-06,],\n",
+      "[-0.44410839676856995,],\n",
+      "[-0.01619231328368187,],\n",
+      "[-0.6207728981971741,],\n",
+      "[-0.2553424835205078,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2075309157371521,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7359861731529236,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.051337532699108124,],\n",
+      "[0.45801839232444763,],\n",
+      "[-0.2828453481197357,],\n",
+      "[1.0112224817276,],\n",
+      "[0.8327498435974121,],\n",
+      "[-0.9400373697280884,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20679493248462677,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0006477718707174063,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.051337532699108124,],\n",
+      "[0.45801839232444763,],\n",
+      "[-0.2828453481197357,],\n",
+      "[1.0112224817276,],\n",
+      "[0.8327498435974121,],\n",
+      "[-0.9400373697280884,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.018812198111846e-08,],\n",
+      "[-6.415891817823649e-08,],\n",
+      "[-1.2449992254914832e-06,],\n",
+      "[5.159418492439727e-07,],\n",
+      "[-0.4563332498073578,],\n",
+      "[-0.017379462718963623,],\n",
+      "[-0.7257575988769531,],\n",
+      "[-0.25460949540138245,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20679493248462677,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.05090273544192314,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.051793865859508514,],\n",
+      "[0.45803576707839966,],\n",
+      "[-0.28211960196495056,],\n",
+      "[1.0114771127700806,],\n",
+      "[0.832808792591095,],\n",
+      "[-0.9400468468666077,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2067440301179886,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.12085144966840744,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.051793865859508514,],\n",
+      "[0.45803576707839966,],\n",
+      "[-0.28211960196495056,],\n",
+      "[1.0114771127700806,],\n",
+      "[0.832808792591095,],\n",
+      "[-0.9400468468666077,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.021567967058218e-07,],\n",
+      "[6.610486593672249e-07,],\n",
+      "[1.3550132280215621e-05,],\n",
+      "[-5.7768365877564065e-06,],\n",
+      "[-0.4679482877254486,],\n",
+      "[-0.018640801310539246,],\n",
+      "[-0.8400259017944336,],\n",
+      "[-0.25456157326698303,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2067440301179886,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6952742338180542,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05226181447505951,],\n",
+      "[0.45805439352989197,],\n",
+      "[-0.2812795639038086,],\n",
+      "[1.0117316246032715,],\n",
+      "[0.8328677415847778,],\n",
+      "[-0.940056324005127,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20743930339813232,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.27307942509651184,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05226181447505951,],\n",
+      "[0.45805439352989197,],\n",
+      "[-0.2812795639038086,],\n",
+      "[1.0117316246032715,],\n",
+      "[0.8328677415847778,],\n",
+      "[-0.940056324005127,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.730494644420105e-07,],\n",
+      "[7.480235240109323e-07,],\n",
+      "[1.6219848475884646e-05,],\n",
+      "[-7.118397206795635e-06,],\n",
+      "[-0.4723864197731018,],\n",
+      "[-0.01924058049917221,],\n",
+      "[-0.8900284767150879,],\n",
+      "[-0.2552568316459656,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20743930339813232,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0451393127441406,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05273419991135597,],\n",
+      "[0.4580736458301544,],\n",
+      "[-0.28038954734802246,],\n",
+      "[1.0119868516921997,],\n",
+      "[0.8329266905784607,],\n",
+      "[-0.9400658011436462,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2084844410419464,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15312789380550385,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05273419991135597,],\n",
+      "[0.4580736458301544,],\n",
+      "[-0.28038954734802246,],\n",
+      "[1.0119868516921997,],\n",
+      "[0.8329266905784607,],\n",
+      "[-0.9400658011436462,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-4.239802251504443e-07,],\n",
+      "[4.2124662513742805e-07,],\n",
+      "[9.669029168435372e-06,],\n",
+      "[-4.369560883787926e-06,],\n",
+      "[-0.4679146409034729,],\n",
+      "[-0.01887437514960766,],\n",
+      "[-0.8474563360214233,],\n",
+      "[-0.2562989890575409,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2084844410419464,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7826323509216309,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05320211499929428,],\n",
+      "[0.45809251070022583,],\n",
+      "[-0.27954208850860596,],\n",
+      "[1.0122431516647339,],\n",
+      "[0.8329856395721436,],\n",
+      "[-0.9400752782821655,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20926707983016968,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0009993934072554111,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.05320211499929428,],\n",
+      "[0.45809251070022583,],\n",
+      "[-0.27954208850860596,],\n",
+      "[1.0122431516647339,],\n",
+      "[0.8329856395721436,],\n",
+      "[-0.9400752782821655,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.5086597688073198e-08,],\n",
+      "[2.5611205600739595e-08,],\n",
+      "[6.219771648829919e-07,],\n",
+      "[-2.893673922699236e-07,],\n",
+      "[-0.4580652415752411,],\n",
+      "[-0.017819751054048538,],\n",
+      "[-0.7442044615745544,],\n",
+      "[-0.25708162784576416,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20926707983016968,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.0632263794541359,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.053660180419683456,],\n",
+      "[0.45811033248901367,],\n",
+      "[-0.2787978947162628,],\n",
+      "[1.012500286102295,],\n",
+      "[0.8330445885658264,],\n",
+      "[-0.9400847554206848,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20933030545711517,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1159423291683197,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.053660180419683456,],\n",
+      "[0.45811033248901367,],\n",
+      "[-0.2787978947162628,],\n",
+      "[1.012500286102295,],\n",
+      "[0.8330445885658264,],\n",
+      "[-0.9400847554206848,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.9790178384937462e-07,],\n",
+      "[-2.0802022504540219e-07,],\n",
+      "[-5.337563834473258e-06,],\n",
+      "[2.5548922621965175e-06,],\n",
+      "[-0.44911742210388184,],\n",
+      "[-0.016750186681747437,],\n",
+      "[-0.6420952677726746,],\n",
+      "[-0.25714486837387085,],\n",
+      "[-0.058973949402570724,],\n",
+      "[0.009499282576143742,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20933030545711517,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6810067892074585,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.054109297692775726,],\n",
+      "[0.45812708139419556,],\n",
+      "[-0.2781558036804199,],\n",
+      "[1.012757420539856,],\n",
+      "[0.8331035375595093,],\n",
+      "[-0.9400942325592041,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20864929258823395,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [8/10], Loss: 0.1159\n",
+      "loss_antes tensor([0.5252150893211365,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6401737332344055,],\n",
+      "[-0.12665215134620667,],\n",
+      "[0.5525429844856262,],\n",
+      "[-1.0006041526794434,],\n",
+      "[0.054109297692775726,],\n",
+      "[0.45812708139419556,],\n",
+      "[-0.2781558036804199,],\n",
+      "[1.012757420539856,],\n",
+      "[0.8331035375595093,],\n",
+      "[-0.9400942325592041,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.29034146666526794,],\n",
+      "[-0.12979142367839813,],\n",
+      "[-0.39259403944015503,],\n",
+      "[0.06027703732252121,],\n",
+      "[-0.2916772961616516,],\n",
+      "[-0.011591554619371891,],\n",
+      "[-0.47371217608451843,],\n",
+      "[-0.18381014466285706,],\n",
+      "[-0.04092193767428398,],\n",
+      "[0.006718185730278492,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20864929258823395,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4494333267211914,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6398833990097046,],\n",
+      "[-0.12652236223220825,],\n",
+      "[0.5529356002807617,],\n",
+      "[-1.000664472579956,],\n",
+      "[0.0544009767472744,],\n",
+      "[0.45813867449760437,],\n",
+      "[-0.2776820957660675,],\n",
+      "[1.0129412412643433,],\n",
+      "[0.83314448595047,],\n",
+      "[-0.9401009678840637,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2071998566389084,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.5130603909492493,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6398833990097046,],\n",
+      "[-0.12652236223220825,],\n",
+      "[0.5529356002807617,],\n",
+      "[-1.000664472579956,],\n",
+      "[0.0544009767472744,],\n",
+      "[0.45813867449760437,],\n",
+      "[-0.2776820957660675,],\n",
+      "[1.0129412412643433,],\n",
+      "[0.83314448595047,],\n",
+      "[-0.9401009678840637,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3086414933204651,],\n",
+      "[-0.1282617300748825,],\n",
+      "[-0.3609689772129059,],\n",
+      "[0.05383531004190445,],\n",
+      "[-0.29301917552948,],\n",
+      "[-0.011008653789758682,],\n",
+      "[-0.4754124879837036,],\n",
+      "[-0.19170111417770386,],\n",
+      "[-0.03848041221499443,],\n",
+      "[0.007541772909462452,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2071998566389084,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.4325627088546753,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6395747661590576,],\n",
+      "[-0.1263941079378128,],\n",
+      "[0.5532965660095215,],\n",
+      "[-1.000718355178833,],\n",
+      "[0.05469399690628052,],\n",
+      "[0.45814967155456543,],\n",
+      "[-0.277206689119339,],\n",
+      "[1.013132929801941,],\n",
+      "[0.8331829905509949,],\n",
+      "[-0.9401085376739502,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20576728880405426,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8470763564109802,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6395747661590576,],\n",
+      "[-0.1263941079378128,],\n",
+      "[0.5532965660095215,],\n",
+      "[-1.000718355178833,],\n",
+      "[0.05469399690628052,],\n",
+      "[0.45814967155456543,],\n",
+      "[-0.277206689119339,],\n",
+      "[1.013132929801941,],\n",
+      "[0.8331829905509949,],\n",
+      "[-0.9401085376739502,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.40696242451667786,],\n",
+      "[-0.15844658017158508,],\n",
+      "[-0.4219711124897003,],\n",
+      "[0.06165241450071335,],\n",
+      "[-0.24647650122642517,],\n",
+      "[-0.008811742067337036,],\n",
+      "[-0.40022605657577515,],\n",
+      "[-0.18298619985580444,],\n",
+      "[-0.03101811744272709,],\n",
+      "[0.007745771668851376,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20576728880405426,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.8407344818115234,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6391677856445312,],\n",
+      "[-0.12623566389083862,],\n",
+      "[0.5537185668945312,],\n",
+      "[-1.0007799863815308,],\n",
+      "[0.05494047328829765,],\n",
+      "[0.4581584930419922,],\n",
+      "[-0.27680647373199463,],\n",
+      "[1.0133159160614014,],\n",
+      "[0.8332139849662781,],\n",
+      "[-0.940116286277771,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20392654836177826,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2541412115097046,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6391677856445312,],\n",
+      "[-0.12623566389083862,],\n",
+      "[0.5537185668945312,],\n",
+      "[-1.0007799863815308,],\n",
+      "[0.05494047328829765,],\n",
+      "[0.4581584930419922,],\n",
+      "[-0.27680647373199463,],\n",
+      "[1.0133159160614014,],\n",
+      "[0.8332139849662781,],\n",
+      "[-0.940116286277771,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.48404619097709656,],\n",
+      "[-0.17862628400325775,],\n",
+      "[-0.45777860283851624,],\n",
+      "[0.06601215153932571,],\n",
+      "[-0.20189327001571655,],\n",
+      "[-0.006706376560032368,],\n",
+      "[-0.32642677426338196,],\n",
+      "[-0.17849981784820557,],\n",
+      "[-0.02484199032187462,],\n",
+      "[0.00815979577600956,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20392654836177826,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.239767074584961,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6386837363243103,],\n",
+      "[-0.12605704367160797,],\n",
+      "[0.5541763305664062,],\n",
+      "[-1.000846028327942,],\n",
+      "[0.0551423653960228,],\n",
+      "[0.4581651985645294,],\n",
+      "[-0.2764800488948822,],\n",
+      "[1.0134943723678589,],\n",
+      "[0.8332388401031494,],\n",
+      "[-0.9401244521141052,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2016867846250534,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.219732403755188,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6386837363243103,],\n",
+      "[-0.12605704367160797,],\n",
+      "[0.5541763305664062,],\n",
+      "[-1.000846028327942,],\n",
+      "[0.0551423653960228,],\n",
+      "[0.4581651985645294,],\n",
+      "[-0.2764800488948822,],\n",
+      "[1.0134943723678589,],\n",
+      "[0.8332388401031494,],\n",
+      "[-0.9401244521141052,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4448535144329071,],\n",
+      "[-0.15774013102054596,],\n",
+      "[-0.3952416181564331,],\n",
+      "[0.05664676055312157,],\n",
+      "[-0.20674479007720947,],\n",
+      "[-0.006176437716931105,],\n",
+      "[-0.33061516284942627,],\n",
+      "[-0.18909814953804016,],\n",
+      "[-0.02628612145781517,],\n",
+      "[0.009008577093482018,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2016867846250534,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.208828926086426,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6382389068603516,],\n",
+      "[-0.1258992999792099,],\n",
+      "[0.5545715689659119,],\n",
+      "[-1.0009026527404785,],\n",
+      "[0.055349111557006836,],\n",
+      "[0.45817136764526367,],\n",
+      "[-0.2761494219303131,],\n",
+      "[1.0136834383010864,],\n",
+      "[0.8332651257514954,],\n",
+      "[-0.9401334524154663,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19947795569896698,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6399304866790771,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6382389068603516,],\n",
+      "[-0.1258992999792099,],\n",
+      "[0.5545715689659119,],\n",
+      "[-1.0009026527404785,],\n",
+      "[0.055349111557006836,],\n",
+      "[0.45817136764526367,],\n",
+      "[-0.2761494219303131,],\n",
+      "[1.0136834383010864,],\n",
+      "[0.8332651257514954,],\n",
+      "[-0.9401334524154663,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.2871573269367218,],\n",
+      "[-0.09916745871305466,],\n",
+      "[-0.24658776819705963,],\n",
+      "[0.035367291420698166,],\n",
+      "[-0.2760685980319977,],\n",
+      "[-0.007437506690621376,],\n",
+      "[-0.4400337338447571,],\n",
+      "[-0.21164989471435547,],\n",
+      "[-0.03564442694187164,],\n",
+      "[0.009998583234846592,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19947795569896698,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.5999113321304321,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6379517316818237,],\n",
+      "[-0.12580013275146484,],\n",
+      "[0.5548181533813477,],\n",
+      "[-1.000938057899475,],\n",
+      "[0.05562518164515495,],\n",
+      "[0.4581788182258606,],\n",
+      "[-0.2757093906402588,],\n",
+      "[1.013895034790039,],\n",
+      "[0.833300769329071,],\n",
+      "[-0.9401434659957886,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19787804782390594,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09958937764167786,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6379517316818237,],\n",
+      "[-0.12580013275146484,],\n",
+      "[0.5548181533813477,],\n",
+      "[-1.000938057899475,],\n",
+      "[0.05562518164515495,],\n",
+      "[0.4581788182258606,],\n",
+      "[-0.2757093906402588,],\n",
+      "[1.013895034790039,],\n",
+      "[0.833300769329071,],\n",
+      "[-0.9401434659957886,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.09702908992767334,],\n",
+      "[-0.03301183506846428,],\n",
+      "[-0.08260808140039444,],\n",
+      "[0.011933136731386185,],\n",
+      "[-0.3834553360939026,],\n",
+      "[-0.009656687267124653,],\n",
+      "[-0.6153037548065186,],\n",
+      "[-0.23536016047000885,],\n",
+      "[-0.04742537438869476,],\n",
+      "[0.010749625973403454,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19787804782390594,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.631155788898468,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378546953201294,],\n",
+      "[-0.12576712667942047,],\n",
+      "[0.5549007654190063,],\n",
+      "[-1.0009499788284302,],\n",
+      "[0.05600863695144653,],\n",
+      "[0.4581884741783142,],\n",
+      "[-0.27509409189224243,],\n",
+      "[1.0141303539276123,],\n",
+      "[0.8333482146263123,],\n",
+      "[-0.9401541948318481,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19724689424037933,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.014798274263739586,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378546953201294,],\n",
+      "[-0.12576712667942047,],\n",
+      "[0.5549007654190063,],\n",
+      "[-1.0009499788284302,],\n",
+      "[0.05600863695144653,],\n",
+      "[0.4581884741783142,],\n",
+      "[-0.27509409189224243,],\n",
+      "[1.0141303539276123,],\n",
+      "[0.8333482146263123,],\n",
+      "[-0.9401541948318481,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.03097974695265293,],\n",
+      "[0.010476267896592617,],\n",
+      "[0.02672095037996769,],\n",
+      "[-0.003908881917595863,],\n",
+      "[-0.4780695140361786,],\n",
+      "[-0.011798305436968803,],\n",
+      "[-0.7740976214408875,],\n",
+      "[-0.25083431601524353,],\n",
+      "[-0.05553070828318596,],\n",
+      "[0.011105521582067013,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19724689424037933,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.24329602718353271,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378856897354126,],\n",
+      "[-0.12577760219573975,],\n",
+      "[0.554874062538147,],\n",
+      "[-1.000946044921875,],\n",
+      "[0.05648670718073845,],\n",
+      "[0.45820027589797974,],\n",
+      "[-0.27432000637054443,],\n",
+      "[1.0143811702728271,],\n",
+      "[0.8334037661552429,],\n",
+      "[-0.9401652812957764,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19749018549919128,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.09242430329322815,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378856897354126,],\n",
+      "[-0.12577760219573975,],\n",
+      "[0.554874062538147,],\n",
+      "[-1.000946044921875,],\n",
+      "[0.05648670718073845,],\n",
+      "[0.45820027589797974,],\n",
+      "[-0.27432000637054443,],\n",
+      "[1.0143811702728271,],\n",
+      "[0.8334037661552429,],\n",
+      "[-0.9401652812957764,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.06239123269915581,],\n",
+      "[0.021116450428962708,],\n",
+      "[0.05549296364188194,],\n",
+      "[-0.008254962973296642,],\n",
+      "[-0.5161557197570801,],\n",
+      "[-0.012958071194589138,],\n",
+      "[-0.8407562971115112,],\n",
+      "[-0.2551253139972687,],\n",
+      "[-0.05759699270129204,],\n",
+      "[0.01115136593580246,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19749018549919128,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6080270409584045,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6379480957984924,],\n",
+      "[-0.12579871714115143,],\n",
+      "[0.5548185706138611,],\n",
+      "[-1.000937819480896,],\n",
+      "[0.05700286477804184,],\n",
+      "[0.4582132399082184,],\n",
+      "[-0.27347925305366516,],\n",
+      "[1.0146362781524658,],\n",
+      "[0.833461344242096,],\n",
+      "[-0.9401764273643494,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19809821248054504,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.028690315783023834,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6379480957984924,],\n",
+      "[-0.12579871714115143,],\n",
+      "[0.5548185706138611,],\n",
+      "[-1.000937819480896,],\n",
+      "[0.05700286477804184,],\n",
+      "[0.4582132399082184,],\n",
+      "[-0.27347925305366516,],\n",
+      "[1.0146362781524658,],\n",
+      "[0.833461344242096,],\n",
+      "[-0.9401764273643494,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.027395091950893402,],\n",
+      "[0.00933935958892107,],\n",
+      "[0.025488940998911858,],\n",
+      "[-0.003867287887260318,],\n",
+      "[-0.4873116612434387,],\n",
+      "[-0.012532385066151619,],\n",
+      "[-0.7920374870300293,],\n",
+      "[-0.25137853622436523,],\n",
+      "[-0.05542152747511864,],\n",
+      "[0.01106435339897871,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19809821248054504,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.33876389265060425,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6379755139350891,],\n",
+      "[-0.12580806016921997,],\n",
+      "[0.5547930598258972,],\n",
+      "[-1.0009340047836304,],\n",
+      "[0.05749017745256424,],\n",
+      "[0.4582257866859436,],\n",
+      "[-0.27268722653388977,],\n",
+      "[1.0148876905441284,],\n",
+      "[0.8335167765617371,],\n",
+      "[-0.9401875138282776,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19843697547912598,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.02574729174375534,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6379755139350891,],\n",
+      "[-0.12580806016921997,],\n",
+      "[0.5547930598258972,],\n",
+      "[-1.0009340047836304,],\n",
+      "[0.05749017745256424,],\n",
+      "[0.4582257866859436,],\n",
+      "[-0.27268722653388977,],\n",
+      "[1.0148876905441284,],\n",
+      "[0.8335167765617371,],\n",
+      "[-0.9401875138282776,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.020084142684936523,],\n",
+      "[-0.006943101994693279,],\n",
+      "[-0.01977458782494068,],\n",
+      "[0.003066800534725189,],\n",
+      "[-0.4199678897857666,],\n",
+      "[-0.010782036930322647,],\n",
+      "[-0.6717844605445862,],\n",
+      "[-0.24530795216560364,],\n",
+      "[-0.05242546647787094,],\n",
+      "[0.010981187224388123,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19843697547912598,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.3209192752838135,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6379554271697998,],\n",
+      "[-0.12580111622810364,],\n",
+      "[0.5548128485679626,],\n",
+      "[-1.0009371042251587,],\n",
+      "[0.057910144329071045,],\n",
+      "[0.45823657512664795,],\n",
+      "[-0.27201545238494873,],\n",
+      "[1.015133023262024,],\n",
+      "[0.8335692286491394,],\n",
+      "[-0.9401984810829163,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19811604917049408,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2049872726202011,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6379554271697998,],\n",
+      "[-0.12580111622810364,],\n",
+      "[0.5548128485679626,],\n",
+      "[-1.0009371042251587,],\n",
+      "[0.057910144329071045,],\n",
+      "[0.45823657512664795,],\n",
+      "[-0.27201545238494873,],\n",
+      "[1.015133023262024,],\n",
+      "[0.8335692286491394,],\n",
+      "[-0.9401984810829163,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04321996867656708,],\n",
+      "[-0.015251239761710167,],\n",
+      "[-0.045440226793289185,],\n",
+      "[0.0072157131507992744,],\n",
+      "[-0.3630475103855133,],\n",
+      "[-0.008964359760284424,],\n",
+      "[-0.5649264454841614,],\n",
+      "[-0.24143877625465393,],\n",
+      "[-0.05095904320478439,],\n",
+      "[0.01095461193472147,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19811604917049408,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9055094122886658,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6379122138023376,],\n",
+      "[-0.12578587234020233,],\n",
+      "[0.5548582673072815,],\n",
+      "[-1.0009443759918213,],\n",
+      "[0.05827319249510765,],\n",
+      "[0.45824554562568665,],\n",
+      "[-0.2714505195617676,],\n",
+      "[1.0153744220733643,],\n",
+      "[0.8336201906204224,],\n",
+      "[-0.9402094483375549,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19721053540706635,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2539920508861542,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6379122138023376,],\n",
+      "[-0.12578587234020233,],\n",
+      "[0.5548582673072815,],\n",
+      "[-1.0009443759918213,],\n",
+      "[0.05827319249510765,],\n",
+      "[0.45824554562568665,],\n",
+      "[-0.2714505195617676,],\n",
+      "[1.0153744220733643,],\n",
+      "[0.8336201906204224,],\n",
+      "[-0.9402094483375549,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.036281343549489975,],\n",
+      "[-0.013136463239789009,],\n",
+      "[-0.04100351780653,],\n",
+      "[0.00667613884434104,],\n",
+      "[-0.3554702401161194,],\n",
+      "[-0.008370691910386086,],\n",
+      "[-0.5459506511688232,],\n",
+      "[-0.2412271946668625,],\n",
+      "[-0.051369015127420425,],\n",
+      "[0.010974306613206863,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19721053540706635,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.0079514980316162,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378759145736694,],\n",
+      "[-0.12577272951602936,],\n",
+      "[0.554899275302887,],\n",
+      "[-1.0009510517120361,],\n",
+      "[0.058628663420677185,],\n",
+      "[0.4582539200782776,],\n",
+      "[-0.2709045708179474,],\n",
+      "[1.015615701675415,],\n",
+      "[0.8336715698242188,],\n",
+      "[-0.9402204155921936,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1962025910615921,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.06881430745124817,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378759145736694,],\n",
+      "[-0.12577272951602936,],\n",
+      "[0.554899275302887,],\n",
+      "[-1.0009510517120361,],\n",
+      "[0.058628663420677185,],\n",
+      "[0.4582539200782776,],\n",
+      "[-0.2709045708179474,],\n",
+      "[1.015615701675415,],\n",
+      "[0.8336715698242188,],\n",
+      "[-0.9402204155921936,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.014126697555184364,],\n",
+      "[-0.005262238904833794,],\n",
+      "[-0.01723797246813774,],\n",
+      "[0.002881099237129092,],\n",
+      "[-0.4030076563358307,],\n",
+      "[-0.009319557808339596,],\n",
+      "[-0.6336930990219116,],\n",
+      "[-0.24354934692382812,],\n",
+      "[-0.052712373435497284,],\n",
+      "[0.01100834459066391,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1962025910615921,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5246496796607971,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378617882728577,],\n",
+      "[-0.12576746940612793,],\n",
+      "[0.554916501045227,],\n",
+      "[-1.0009539127349854,],\n",
+      "[0.05903167277574539,],\n",
+      "[0.45826324820518494,],\n",
+      "[-0.2702708840370178,],\n",
+      "[1.0158592462539673,],\n",
+      "[0.8337242603302002,],\n",
+      "[-0.940231442451477,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19567793607711792,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.020092038437724113,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378617882728577,],\n",
+      "[-0.12576746940612793,],\n",
+      "[0.554916501045227,],\n",
+      "[-1.0009539127349854,],\n",
+      "[0.05903167277574539,],\n",
+      "[0.45826324820518494,],\n",
+      "[-0.2702708840370178,],\n",
+      "[1.0158592462539673,],\n",
+      "[0.8337242603302002,],\n",
+      "[-0.940231442451477,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005678274668753147,],\n",
+      "[0.0021763229742646217,],\n",
+      "[0.007501383312046528,],\n",
+      "[-0.001288248342461884,],\n",
+      "[-0.4770161211490631,],\n",
+      "[-0.011193862184882164,],\n",
+      "[-0.7805507183074951,],\n",
+      "[-0.2464275062084198,],\n",
+      "[-0.05390527844429016,],\n",
+      "[0.01103267539292574,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19567793607711792,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.28349289298057556,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378674507141113,],\n",
+      "[-0.12576964497566223,],\n",
+      "[0.5549089908599854,],\n",
+      "[-1.0009526014328003,],\n",
+      "[0.05950868874788284,],\n",
+      "[0.4582744538784027,],\n",
+      "[-0.2694903314113617,],\n",
+      "[1.0161056518554688,],\n",
+      "[0.8337781429290771,],\n",
+      "[-0.9402424693107605,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19596143066883087,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22612708806991577,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378674507141113,],\n",
+      "[-0.12576964497566223,],\n",
+      "[0.5549089908599854,],\n",
+      "[-1.0009526014328003,],\n",
+      "[0.05950868874788284,],\n",
+      "[0.4582744538784027,],\n",
+      "[-0.2694903314113617,],\n",
+      "[1.0161056518554688,],\n",
+      "[0.8337781429290771,],\n",
+      "[-0.9402424693107605,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.014115522615611553,],\n",
+      "[0.005558785516768694,],\n",
+      "[0.020220128819346428,],\n",
+      "[-0.0035709007643163204,],\n",
+      "[-0.5342218279838562,],\n",
+      "[-0.013020236976444721,],\n",
+      "[-0.9019284248352051,],\n",
+      "[-0.24844273924827576,],\n",
+      "[-0.0544106587767601,],\n",
+      "[0.0110402163118124,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19596143066883087,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9510572552680969,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378815770149231,],\n",
+      "[-0.12577520310878754,],\n",
+      "[0.5548887848854065,],\n",
+      "[-1.0009490251541138,],\n",
+      "[0.060042910277843475,],\n",
+      "[0.45828747749328613,],\n",
+      "[-0.2685883939266205,],\n",
+      "[1.0163540840148926,],\n",
+      "[0.8338325619697571,],\n",
+      "[-0.940253496170044,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.196912482380867,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2981552481651306,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378815770149231,],\n",
+      "[-0.12577520310878754,],\n",
+      "[0.5548887848854065,],\n",
+      "[-1.0009490251541138,],\n",
+      "[0.060042910277843475,],\n",
+      "[0.45828747749328613,],\n",
+      "[-0.2685883939266205,],\n",
+      "[1.0163540840148926,],\n",
+      "[0.8338325619697571,],\n",
+      "[-0.940253496170044,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.011975232511758804,],\n",
+      "[0.00483947666361928,],\n",
+      "[0.018620101734995842,],\n",
+      "[-0.0033836516086012125,],\n",
+      "[-0.5435363054275513,],\n",
+      "[-0.01383763924241066,],\n",
+      "[-0.9277629852294922,],\n",
+      "[-0.24917082488536835,],\n",
+      "[-0.05428843945264816,],\n",
+      "[0.011035851202905178,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.196912482380867,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0920718908309937,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.637893557548523,],\n",
+      "[-0.12578004598617554,],\n",
+      "[0.5548701882362366,],\n",
+      "[-1.0009456872940063,],\n",
+      "[0.06058644503355026,],\n",
+      "[0.458301305770874,],\n",
+      "[-0.26766061782836914,],\n",
+      "[1.0166032314300537,],\n",
+      "[0.8338868618011475,],\n",
+      "[-0.9402645230293274,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1980045586824417,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10100452601909637,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.637893557548523,],\n",
+      "[-0.12578004598617554,],\n",
+      "[0.5548701882362366,],\n",
+      "[-1.0009456872940063,],\n",
+      "[0.06058644503355026,],\n",
+      "[0.458301305770874,],\n",
+      "[-0.26766061782836914,],\n",
+      "[1.0166032314300537,],\n",
+      "[0.8338868618011475,],\n",
+      "[-0.9402645230293274,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005137114319950342,],\n",
+      "[0.002130647888407111,],\n",
+      "[0.008677837438881397,],\n",
+      "[-0.0016232752241194248,],\n",
+      "[-0.5036482810974121,],\n",
+      "[-0.013115251436829567,],\n",
+      "[-0.8452959656715393,],\n",
+      "[-0.24889495968818665,],\n",
+      "[-0.053895313292741776,],\n",
+      "[0.011028127744793892,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1980045586824417,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6356239318847656,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378986835479736,],\n",
+      "[-0.12578217685222626,],\n",
+      "[0.5548614859580994,],\n",
+      "[-1.0009440183639526,],\n",
+      "[0.06109009310603142,],\n",
+      "[0.4583144187927246,],\n",
+      "[-0.26681533455848694,],\n",
+      "[1.0168521404266357,],\n",
+      "[0.8339407444000244,],\n",
+      "[-0.9402755498886108,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19864018261432648,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0042244563810527325,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378986835479736,],\n",
+      "[-0.12578217685222626,],\n",
+      "[0.5548614859580994,],\n",
+      "[-1.0009440183639526,],\n",
+      "[0.06109009310603142,],\n",
+      "[0.4583144187927246,],\n",
+      "[-0.26681533455848694,],\n",
+      "[1.0168521404266357,],\n",
+      "[0.8339407444000244,],\n",
+      "[-0.9402755498886108,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0007727343472652137,],\n",
+      "[-0.0003294905473012477,],\n",
+      "[-0.0014193907845765352,],\n",
+      "[0.0002733402361627668,],\n",
+      "[-0.4421597719192505,],\n",
+      "[-0.011296363547444344,],\n",
+      "[-0.7070976495742798,],\n",
+      "[-0.248138427734375,],\n",
+      "[-0.053559307008981705,],\n",
+      "[0.011022599413990974,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19864018261432648,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.12999172508716583,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378979086875916,],\n",
+      "[-0.12578184902668,],\n",
+      "[0.554862916469574,],\n",
+      "[-1.0009442567825317,],\n",
+      "[0.06153225153684616,],\n",
+      "[0.45832571387290955,],\n",
+      "[-0.2661082446575165,],\n",
+      "[1.0171003341674805,],\n",
+      "[0.8339943289756775,],\n",
+      "[-0.9402865767478943,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19851018488407135,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13375692069530487,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378979086875916,],\n",
+      "[-0.12578184902668,],\n",
+      "[0.554862916469574,],\n",
+      "[-1.0009442567825317,],\n",
+      "[0.06153225153684616,],\n",
+      "[0.45832571387290955,],\n",
+      "[-0.2661082446575165,],\n",
+      "[1.0171003341674805,],\n",
+      "[0.8339943289756775,],\n",
+      "[-0.9402865767478943,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0031929872930049896,],\n",
+      "[-0.0014031316386535764,],\n",
+      "[-0.0063826339319348335,],\n",
+      "[0.0012652251170948148,],\n",
+      "[-0.3975526690483093,],\n",
+      "[-0.009658845141530037,],\n",
+      "[-0.5988432765007019,],\n",
+      "[-0.24732527136802673,],\n",
+      "[-0.053422506898641586,],\n",
+      "[0.011020886711776257,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19851018488407135,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7314562797546387,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946900367737,],\n",
+      "[-0.12578044831752777,],\n",
+      "[0.5548692941665649,],\n",
+      "[-1.0009455680847168,],\n",
+      "[0.06192980334162712,],\n",
+      "[0.45833536982536316,],\n",
+      "[-0.26550939679145813,],\n",
+      "[1.0173476934432983,],\n",
+      "[0.8340477347373962,],\n",
+      "[-0.9402976036071777,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19777873158454895,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15890003740787506,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946900367737,],\n",
+      "[-0.12578044831752777,],\n",
+      "[0.5548692941665649,],\n",
+      "[-1.0009455680847168,],\n",
+      "[0.06192980334162712,],\n",
+      "[0.45833536982536316,],\n",
+      "[-0.26550939679145813,],\n",
+      "[1.0173476934432983,],\n",
+      "[0.8340477347373962,],\n",
+      "[-0.9402976036071777,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.002552715362980962,],\n",
+      "[-0.0011582404840737581,],\n",
+      "[-0.0055561731569468975,],\n",
+      "[0.0011335514718666673,],\n",
+      "[-0.3951376974582672,],\n",
+      "[-0.00930687878280878,],\n",
+      "[-0.5870192646980286,],\n",
+      "[-0.24671360850334167,],\n",
+      "[-0.053457509726285934,],\n",
+      "[0.011021758429706097,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19777873158454895,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7972460389137268,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378921270370483,],\n",
+      "[-0.12577928602695465,],\n",
+      "[0.554874837398529,],\n",
+      "[-1.0009467601776123,],\n",
+      "[0.06232494115829468,],\n",
+      "[0.4583446681499481,],\n",
+      "[-0.2649223804473877,],\n",
+      "[1.0175944566726685,],\n",
+      "[0.8341012001037598,],\n",
+      "[-0.9403086304664612,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1969814896583557,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.020001502707600594,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378921270370483,],\n",
+      "[-0.12577928602695465,],\n",
+      "[0.554874837398529,],\n",
+      "[-1.0009467601776123,],\n",
+      "[0.06232494115829468,],\n",
+      "[0.4583446681499481,],\n",
+      "[-0.2649223804473877,],\n",
+      "[1.0175944566726685,],\n",
+      "[0.8341012001037598,],\n",
+      "[-0.9403086304664612,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0006638472550548613,],\n",
+      "[-0.00031107632094062865,],\n",
+      "[-0.0015735123306512833,],\n",
+      "[0.0003304042329546064,],\n",
+      "[-0.4326896667480469,],\n",
+      "[-0.010377036407589912,],\n",
+      "[-0.6790217161178589,],\n",
+      "[-0.24646632373332977,],\n",
+      "[-0.0535602830350399,],\n",
+      "[0.01102322619408369,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1969814896583557,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.28285351395606995,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378914713859558,],\n",
+      "[-0.12577897310256958,],\n",
+      "[0.5548763871192932,],\n",
+      "[-1.000947117805481,],\n",
+      "[0.06275763362646103,],\n",
+      "[0.45835503935813904,],\n",
+      "[-0.26424336433410645,],\n",
+      "[1.0178409814834595,],\n",
+      "[0.8341547846794128,],\n",
+      "[-0.9403196573257446,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1966986358165741,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.06309904903173447,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378914713859558,],\n",
+      "[-0.12577897310256958,],\n",
+      "[0.5548763871192932,],\n",
+      "[-1.000947117805481,],\n",
+      "[0.06275763362646103,],\n",
+      "[0.45835503935813904,],\n",
+      "[-0.26424336433410645,],\n",
+      "[1.0178409814834595,],\n",
+      "[0.8341547846794128,],\n",
+      "[-0.9403196573257446,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.000863912224303931,],\n",
+      "[0.00041749730007722974,],\n",
+      "[0.002229338278993964,],\n",
+      "[-0.0004819472087547183,],\n",
+      "[-0.4845321476459503,],\n",
+      "[-0.012169700115919113,],\n",
+      "[-0.8190448880195618,],\n",
+      "[-0.24670588970184326,],\n",
+      "[-0.053642306476831436,],\n",
+      "[0.011024290695786476,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1966986358165741,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5023902654647827,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378923058509827,],\n",
+      "[-0.125779390335083,],\n",
+      "[0.5548741817474365,],\n",
+      "[-1.0009466409683228,],\n",
+      "[0.06324216723442078,],\n",
+      "[0.4583671987056732,],\n",
+      "[-0.2634243071079254,],\n",
+      "[1.0180877447128296,],\n",
+      "[0.8342084288597107,],\n",
+      "[-0.9403306841850281,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19720102846622467,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2933945059776306,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378923058509827,],\n",
+      "[-0.125779390335083,],\n",
+      "[0.5548741817474365,],\n",
+      "[-1.0009466409683228,],\n",
+      "[0.06324216723442078,],\n",
+      "[0.4583671987056732,],\n",
+      "[-0.2634243071079254,],\n",
+      "[1.0180877447128296,],\n",
+      "[0.8342084288597107,],\n",
+      "[-0.9403306841850281,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0013645516010001302,],\n",
+      "[0.0006786022568121552,],\n",
+      "[0.0038316415157169104,],\n",
+      "[-0.0008532252395525575,],\n",
+      "[-0.5191493630409241,],\n",
+      "[-0.013693171553313732,],\n",
+      "[-0.9219668507575989,],\n",
+      "[-0.2474275827407837,],\n",
+      "[-0.05366925522685051,],\n",
+      "[0.01102452166378498,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19720102846622467,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.083318829536438,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378936767578125,],\n",
+      "[-0.12578007578849792,],\n",
+      "[0.5548703670501709,],\n",
+      "[-1.000945806503296,],\n",
+      "[0.06376131623983383,],\n",
+      "[0.45838087797164917,],\n",
+      "[-0.2625023424625397,],\n",
+      "[1.018335223197937,],\n",
+      "[0.8342620730400085,],\n",
+      "[-0.9403417110443115,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1982843428850174,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3067382872104645,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378936767578125,],\n",
+      "[-0.12578007578849792,],\n",
+      "[0.5548703670501709,],\n",
+      "[-1.000945806503296,],\n",
+      "[0.06376131623983383,],\n",
+      "[0.45838087797164917,],\n",
+      "[-0.2625023424625397,],\n",
+      "[1.018335223197937,],\n",
+      "[0.8342620730400085,],\n",
+      "[-0.9403417110443115,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0010217769304290414,],\n",
+      "[0.0005220645689405501,],\n",
+      "[0.0031207441352307796,],\n",
+      "[-0.0007160806562751532,],\n",
+      "[-0.5177974104881287,],\n",
+      "[-0.014099902473390102,],\n",
+      "[-0.9255726337432861,],\n",
+      "[-0.24841339886188507,],\n",
+      "[-0.053651612251996994,],\n",
+      "[0.01102417055517435,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1982843428850174,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.1076804399490356,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946900367737,],\n",
+      "[-0.1257805973291397,],\n",
+      "[0.5548672676086426,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.06427911669015884,],\n",
+      "[0.45839497447013855,],\n",
+      "[-0.261576771736145,],\n",
+      "[1.0185836553573608,],\n",
+      "[0.8343157172203064,],\n",
+      "[-0.940352737903595,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19939202070236206,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07849416136741638,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946900367737,],\n",
+      "[-0.1257805973291397,],\n",
+      "[0.5548672676086426,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.06427911669015884,],\n",
+      "[0.45839497447013855,],\n",
+      "[-0.261576771736145,],\n",
+      "[1.0185836553573608,],\n",
+      "[0.8343157172203064,],\n",
+      "[-0.940352737903595,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00037844362668693066,],\n",
+      "[0.00019864810747094452,],\n",
+      "[0.0012571269180625677,],\n",
+      "[-0.0002972545917145908,],\n",
+      "[-0.48404818773269653,],\n",
+      "[-0.013129444792866707,],\n",
+      "[-0.8289672136306763,],\n",
+      "[-0.24926303327083588,],\n",
+      "[-0.05361833795905113,],\n",
+      "[0.011023747734725475,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19939202070236206,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5603367686271667,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378950476646423,],\n",
+      "[-0.12578079104423523,],\n",
+      "[0.5548660159111023,],\n",
+      "[-1.0009448528289795,],\n",
+      "[0.06476316601037979,],\n",
+      "[0.4584081172943115,],\n",
+      "[-0.2607477903366089,],\n",
+      "[1.0188329219818115,],\n",
+      "[0.8343693614006042,],\n",
+      "[-0.9403637647628784,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1999523639678955,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.012914542108774185,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378950476646423,],\n",
+      "[-0.12578079104423523,],\n",
+      "[0.5548660159111023,],\n",
+      "[-1.0009448528289795,],\n",
+      "[0.06476316601037979,],\n",
+      "[0.4584081172943115,],\n",
+      "[-0.2607477903366089,],\n",
+      "[1.0188329219818115,],\n",
+      "[0.8343693614006042,],\n",
+      "[-0.9403637647628784,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00011236789578106254,],\n",
+      "[-6.0687310906359926e-05,],\n",
+      "[-0.0004061004729010165,],\n",
+      "[9.89175823633559e-05,],\n",
+      "[-0.4401867985725403,],\n",
+      "[-0.011412229388952255,],\n",
+      "[-0.6919829249382019,],\n",
+      "[-0.24959169328212738,],\n",
+      "[-0.05359374359250069,],\n",
+      "[0.011023450642824173,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1999523639678955,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.22728443145751953,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378949284553528,],\n",
+      "[-0.12578073143959045,],\n",
+      "[0.5548664331436157,],\n",
+      "[-1.000944972038269,],\n",
+      "[0.06520335376262665,],\n",
+      "[0.458419531583786,],\n",
+      "[-0.260055810213089,],\n",
+      "[1.0190825462341309,],\n",
+      "[0.8344229459762573,],\n",
+      "[-0.9403747916221619,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19972507655620575,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1514282077550888,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378949284553528,],\n",
+      "[-0.12578073143959045,],\n",
+      "[0.5548664331436157,],\n",
+      "[-1.000944972038269,],\n",
+      "[0.06520335376262665,],\n",
+      "[0.458419531583786,],\n",
+      "[-0.260055810213089,],\n",
+      "[1.0190825462341309,],\n",
+      "[0.8344229459762573,],\n",
+      "[-0.9403747916221619,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0002816148044075817,],\n",
+      "[-0.00015682991943322122,],\n",
+      "[-0.0011078392853960395,],\n",
+      "[0.0002778225752990693,],\n",
+      "[-0.4126925766468048,],\n",
+      "[-0.010105877183377743,],\n",
+      "[-0.5971018075942993,],\n",
+      "[-0.24926628172397614,],\n",
+      "[-0.05358506739139557,],\n",
+      "[0.011023455299437046,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19972507655620575,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7782756686210632,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548675656318665,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.06561604887247086,],\n",
+      "[0.45842963457107544,],\n",
+      "[-0.2594587206840515,],\n",
+      "[1.0193318128585815,],\n",
+      "[0.8344765305519104,],\n",
+      "[-0.9403858184814453,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19894680380821228,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1442994922399521,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548675656318665,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.06561604887247086,],\n",
+      "[0.45842963457107544,],\n",
+      "[-0.2594587206840515,],\n",
+      "[1.0193318128585815,],\n",
+      "[0.8344765305519104,],\n",
+      "[-0.9403858184814453,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00020118270185776055,],\n",
+      "[-0.000115700822789222,],\n",
+      "[-0.0008617761777713895,],\n",
+      "[0.00022241177794057876,],\n",
+      "[-0.41572993993759155,],\n",
+      "[-0.010021544061601162,],\n",
+      "[-0.6002852916717529,],\n",
+      "[-0.24852438271045685,],\n",
+      "[-0.053588900715112686,],\n",
+      "[0.011023466475307941,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19894680380821228,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.759734570980072,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378944516181946,],\n",
+      "[-0.12578044831752777,],\n",
+      "[0.5548684000968933,],\n",
+      "[-1.0009454488754272,],\n",
+      "[0.06603177636861801,],\n",
+      "[0.4584396481513977,],\n",
+      "[-0.25885844230651855,],\n",
+      "[1.019580364227295,],\n",
+      "[0.8345301151275635,],\n",
+      "[-0.9403968453407288,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19818706810474396,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.008552192710340023,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378944516181946,],\n",
+      "[-0.12578044831752777,],\n",
+      "[0.5548684000968933,],\n",
+      "[-1.0009454488754272,],\n",
+      "[0.06603177636861801,],\n",
+      "[0.4584396481513977,],\n",
+      "[-0.25885844230651855,],\n",
+      "[1.019580364227295,],\n",
+      "[0.8345301151275635,],\n",
+      "[-0.9403968453407288,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.584088699426502e-05,],\n",
+      "[-2.1284775357344188e-05,],\n",
+      "[-0.00016716554819140583,],\n",
+      "[4.4399595935828984e-05,],\n",
+      "[-0.44381606578826904,],\n",
+      "[-0.011107919737696648,],\n",
+      "[-0.6986309885978699,],\n",
+      "[-0.24786049127578735,],\n",
+      "[-0.05359691008925438,],\n",
+      "[0.011023567058146,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19818706810474396,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.1849561631679535,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378943920135498,],\n",
+      "[-0.12578043341636658,],\n",
+      "[0.5548685789108276,],\n",
+      "[-1.0009454488754272,],\n",
+      "[0.06647559255361557,],\n",
+      "[0.4584507644176483,],\n",
+      "[-0.2581598162651062,],\n",
+      "[1.019828200340271,],\n",
+      "[0.8345836997032166,],\n",
+      "[-0.9404078722000122,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19800211489200592,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0883575826883316,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378943920135498,],\n",
+      "[-0.12578043341636658,],\n",
+      "[0.5548685789108276,],\n",
+      "[-1.0009454488754272,],\n",
+      "[0.06647559255361557,],\n",
+      "[0.4584507644176483,],\n",
+      "[-0.2581598162651062,],\n",
+      "[1.019828200340271,],\n",
+      "[0.8345836997032166,],\n",
+      "[-0.9404078722000122,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.430147863691673e-05,],\n",
+      "[5.1615617849165574e-05,],\n",
+      "[0.00042792301974259317,],\n",
+      "[-0.00011701700714183971,],\n",
+      "[-0.47762376070022583,],\n",
+      "[-0.012654246762394905,],\n",
+      "[-0.8313384056091309,],\n",
+      "[-0.24775907397270203,],\n",
+      "[-0.05360283702611923,],\n",
+      "[0.011023567989468575,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19800211489200592,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5945003032684326,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378944516181946,],\n",
+      "[-0.12578047811985016,],\n",
+      "[0.5548681616783142,],\n",
+      "[-1.0009453296661377,],\n",
+      "[0.06695321947336197,],\n",
+      "[0.4584634304046631,],\n",
+      "[-0.25732848048210144,],\n",
+      "[1.0200759172439575,],\n",
+      "[0.8346372842788696,],\n",
+      "[-0.9404188990592957,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19859661161899567,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.30375853180885315,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378944516181946,],\n",
+      "[-0.12578047811985016,],\n",
+      "[0.5548681616783142,],\n",
+      "[-1.0009453296661377,],\n",
+      "[0.06695321947336197,],\n",
+      "[0.4584634304046631,],\n",
+      "[-0.25732848048210144,],\n",
+      "[1.0200759172439575,],\n",
+      "[0.8346372842788696,],\n",
+      "[-0.9404188990592957,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0001143785921158269,],\n",
+      "[7.204661233117804e-05,],\n",
+      "[0.0006314891506917775,],\n",
+      "[-0.0001778923033270985,],\n",
+      "[-0.49690622091293335,],\n",
+      "[-0.013777084648609161,],\n",
+      "[-0.9160774946212769,],\n",
+      "[-0.24838164448738098,],\n",
+      "[-0.053604401648044586,],\n",
+      "[0.01102356519550085,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19859661161899567,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.1022857427597046,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378945708274841,],\n",
+      "[-0.12578055262565613,],\n",
+      "[0.5548675060272217,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.0674501284956932,],\n",
+      "[0.4584771990776062,],\n",
+      "[-0.25641241669654846,],\n",
+      "[1.0203243494033813,],\n",
+      "[0.8346908688545227,],\n",
+      "[-0.9404299259185791,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19969889521598816,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2635402977466583,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378945708274841,],\n",
+      "[-0.12578055262565613,],\n",
+      "[0.5548675060272217,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.0674501284956932,],\n",
+      "[0.4584771990776062,],\n",
+      "[-0.25641241669654846,],\n",
+      "[1.0203243494033813,],\n",
+      "[0.8346908688545227,],\n",
+      "[-0.9404299259185791,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.795832061674446e-05,],\n",
+      "[5.044855788582936e-05,],\n",
+      "[0.00046794014633633196,],\n",
+      "[-0.00013585036504082382,],\n",
+      "[-0.491818368434906,],\n",
+      "[-0.013847197405993938,],\n",
+      "[-0.9013513922691345,],\n",
+      "[-0.24946121871471405,],\n",
+      "[-0.053602639585733414,],\n",
+      "[0.011023561470210552,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19969889521598816,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0267243385314941,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.1257805973291397,],\n",
+      "[0.5548670291900635,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.06794194877147675,],\n",
+      "[0.4584910571575165,],\n",
+      "[-0.255511075258255,],\n",
+      "[1.0205738544464111,],\n",
+      "[0.8347444534301758,],\n",
+      "[-0.9404409527778625,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20072561502456665,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04205916076898575,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.1257805973291397,],\n",
+      "[0.5548670291900635,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.06794194877147675,],\n",
+      "[0.4584910571575165,],\n",
+      "[-0.255511075258255,],\n",
+      "[1.0205738544464111,],\n",
+      "[0.8347444534301758,],\n",
+      "[-0.9404409527778625,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.2788675778429024e-05,],\n",
+      "[1.5153278582147323e-05,],\n",
+      "[0.00014871812891215086,],\n",
+      "[-4.449260450201109e-05,],\n",
+      "[-0.46739327907562256,],\n",
+      "[-0.012839937582612038,],\n",
+      "[-0.7987092137336731,],\n",
+      "[-0.2504429817199707,],\n",
+      "[-0.05360009893774986,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20072561502456665,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4101664125919342,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.1257806122303009,],\n",
+      "[0.5548669099807739,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.06840933859348297,],\n",
+      "[0.4585039019584656,],\n",
+      "[-0.25471237301826477,],\n",
+      "[1.0208243131637573,],\n",
+      "[0.8347980380058289,],\n",
+      "[-0.940451979637146,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.2011357843875885,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03549080342054367,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.1257806122303009,],\n",
+      "[0.5548669099807739,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.06840933859348297,],\n",
+      "[0.4585039019584656,],\n",
+      "[-0.25471237301826477,],\n",
+      "[1.0208243131637573,],\n",
+      "[0.8347980380058289,],\n",
+      "[-0.940451979637146,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.5317522411351092e-05,],\n",
+      "[-1.0483588994247839e-05,],\n",
+      "[-0.0001087304117390886,],\n",
+      "[3.350519182276912e-05,],\n",
+      "[-0.43952855467796326,],\n",
+      "[-0.011417422443628311,],\n",
+      "[-0.671696662902832,],\n",
+      "[-0.25081729888916016,],\n",
+      "[-0.053598254919052124,],\n",
+      "[0.011023558676242828,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.2011357843875885,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.37677982449531555,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.1257805973291397,],\n",
+      "[0.5548670291900635,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.0688488706946373,],\n",
+      "[0.45851531624794006,],\n",
+      "[-0.2540406882762909,],\n",
+      "[1.0210751295089722,],\n",
+      "[0.8348516225814819,],\n",
+      "[-0.9404630064964294,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20075900852680206,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.18495729565620422,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.1257805973291397,],\n",
+      "[0.5548670291900635,],\n",
+      "[-1.0009450912475586,],\n",
+      "[0.0688488706946373,],\n",
+      "[0.45851531624794006,],\n",
+      "[-0.2540406882762909,],\n",
+      "[1.0210751295089722,],\n",
+      "[0.8348516225814819,],\n",
+      "[-0.9404630064964294,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.5585852199583314e-05,],\n",
+      "[-1.8063286916003563e-05,],\n",
+      "[-0.0001976727944565937,],\n",
+      "[6.269945879466832e-05,],\n",
+      "[-0.42467960715293884,],\n",
+      "[-0.010498099029064178,],\n",
+      "[-0.5947608947753906,],\n",
+      "[-0.2504286468029022,],\n",
+      "[-0.05359787121415138,],\n",
+      "[0.011023558676242828,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20075900852680206,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.860132098197937,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578058242797852,],\n",
+      "[0.5548672080039978,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.06927355378866196,],\n",
+      "[0.45852580666542053,],\n",
+      "[-0.25344592332839966,],\n",
+      "[1.0213255882263184,],\n",
+      "[0.834905207157135,],\n",
+      "[-0.9404740333557129,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1998988837003708,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14065831899642944,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578058242797852,],\n",
+      "[0.5548672080039978,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.06927355378866196,],\n",
+      "[0.45852580666542053,],\n",
+      "[-0.25344592332839966,],\n",
+      "[1.0213255882263184,],\n",
+      "[0.834905207157135,],\n",
+      "[-0.9404740333557129,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.6325877368217334e-05,],\n",
+      "[-1.1904831808351446e-05,],\n",
+      "[-0.0001373328414047137,],\n",
+      "[4.481980431592092e-05,],\n",
+      "[-0.42972415685653687,],\n",
+      "[-0.010603990405797958,],\n",
+      "[-0.6118912696838379,],\n",
+      "[-0.24957607686519623,],\n",
+      "[-0.05359822139143944,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1998988837003708,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.750088632106781,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548673272132874,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.06970328092575073,],\n",
+      "[0.45853641629219055,],\n",
+      "[-0.2528340220451355,],\n",
+      "[1.0215752124786377,],\n",
+      "[0.8349587917327881,],\n",
+      "[-0.9404850602149963,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19914878904819489,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0034336531534790993,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548673272132874,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.06970328092575073,],\n",
+      "[0.45853641629219055,],\n",
+      "[-0.2528340220451355,],\n",
+      "[1.0215752124786377,],\n",
+      "[0.8349587917327881,],\n",
+      "[-0.9404850602149963,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.866383058768406e-06,],\n",
+      "[-1.4052793630980887e-06,],\n",
+      "[-1.7092499547288753e-05,],\n",
+      "[5.7400516197958495e-06,],\n",
+      "[-0.4490208327770233,],\n",
+      "[-0.011572657153010368,],\n",
+      "[-0.7122207283973694,],\n",
+      "[-0.24884246289730072,],\n",
+      "[-0.053598951548337936,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19914878904819489,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.11719468981027603,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548673272132874,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07015230506658554,],\n",
+      "[0.458547979593277,],\n",
+      "[-0.25212180614471436,],\n",
+      "[1.0218240022659302,],\n",
+      "[0.8350123763084412,],\n",
+      "[-0.9404960870742798,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19903159141540527,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10566990822553635,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548673272132874,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07015230506658554,],\n",
+      "[0.458547979593277,],\n",
+      "[-0.25212180614471436,],\n",
+      "[1.0218240022659302,],\n",
+      "[0.8350123763084412,],\n",
+      "[-0.9404960870742798,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.5757693593914155e-06,],\n",
+      "[5.879613581782905e-06,],\n",
+      "[7.549380825366825e-05,],\n",
+      "[-2.610118463053368e-05,],\n",
+      "[-0.4695676565170288,],\n",
+      "[-0.012773017399013042,],\n",
+      "[-0.8332338333129883,],\n",
+      "[-0.24873869121074677,],\n",
+      "[-0.05359932407736778,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19903159141540527,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.650137722492218,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548672676086426,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07062187045812607,],\n",
+      "[0.4585607647895813,],\n",
+      "[-0.2512885630130768,],\n",
+      "[1.0220727920532227,],\n",
+      "[0.8350659608840942,],\n",
+      "[-0.9405071139335632,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19968172907829285,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2925109565258026,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548672676086426,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07062187045812607,],\n",
+      "[0.4585607647895813,],\n",
+      "[-0.2512885630130768,],\n",
+      "[1.0220727920532227,],\n",
+      "[0.8350659608840942,],\n",
+      "[-0.9405071139335632,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-9.22252093005227e-06,],\n",
+      "[7.362404630839592e-06,],\n",
+      "[9.993887215387076e-05,],\n",
+      "[-3.559580000001006e-05,],\n",
+      "[-0.4793745279312134,],\n",
+      "[-0.013510018587112427,],\n",
+      "[-0.8984279036521912,],\n",
+      "[-0.24939176440238953,],\n",
+      "[-0.05359933152794838,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19968172907829285,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0816856622695923,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07110124826431274,],\n",
+      "[0.4585742652416229,],\n",
+      "[-0.2503901422023773,],\n",
+      "[1.022322177886963,],\n",
+      "[0.8351195454597473,],\n",
+      "[-0.9405181407928467,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20076341927051544,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20762130618095398,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07110124826431274,],\n",
+      "[0.4585742652416229,],\n",
+      "[-0.2503901422023773,],\n",
+      "[1.022322177886963,],\n",
+      "[0.8351195454597473,],\n",
+      "[-0.9405181407928467,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.685157248080941e-06,],\n",
+      "[4.662751507567009e-06,],\n",
+      "[6.696818309137598e-05,],\n",
+      "[-2.4581224352004938e-05,],\n",
+      "[-0.4739006757736206,],\n",
+      "[-0.013379535637795925,],\n",
+      "[-0.869003415107727,],\n",
+      "[-0.250468909740448,],\n",
+      "[-0.053599338978528976,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20076341927051544,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9113088846206665,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548670887947083,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07157514989376068,],\n",
+      "[0.458587646484375,],\n",
+      "[-0.2495211362838745,],\n",
+      "[1.022572636604309,],\n",
+      "[0.8351731300354004,],\n",
+      "[-0.9405291676521301,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20167472958564758,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.014191588386893272,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548670887947083,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07157514989376068,],\n",
+      "[0.458587646484375,],\n",
+      "[-0.2495211362838745,],\n",
+      "[1.022572636604309,],\n",
+      "[0.8351731300354004,],\n",
+      "[-0.9405291676521301,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.0875472753468785e-06,],\n",
+      "[9.168089718514238e-07,],\n",
+      "[1.3927522559242789e-05,],\n",
+      "[-5.26764415553771e-06,],\n",
+      "[-0.4577712118625641,],\n",
+      "[-0.012494738213717937,],\n",
+      "[-0.7674652934074402,],\n",
+      "[-0.251374214887619,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20167472958564758,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2382565438747406,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548670887947083,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07203292101621628,],\n",
+      "[0.45860013365745544,],\n",
+      "[-0.24875366687774658,],\n",
+      "[1.0228240489959717,],\n",
+      "[0.8352267146110535,],\n",
+      "[-0.9405401945114136,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.201912984251976,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07114756852388382,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548670887947083,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07203292101621628,],\n",
+      "[0.45860013365745544,],\n",
+      "[-0.24875366687774658,],\n",
+      "[1.0228240489959717,],\n",
+      "[0.8352267146110535,],\n",
+      "[-0.9405401945114136,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.78171990228293e-06,],\n",
+      "[-1.5466874856429058e-06,],\n",
+      "[-2.4820154067128897e-05,],\n",
+      "[9.667385711509269e-06,],\n",
+      "[-0.44138333201408386,],\n",
+      "[-0.011433770880103111,],\n",
+      "[-0.6561653017997742,],\n",
+      "[-0.2516064941883087,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.201912984251976,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5334702730178833,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548670887947083,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07247430086135864,],\n",
+      "[0.4586115777492523,],\n",
+      "[-0.2480974942445755,],\n",
+      "[1.0230756998062134,],\n",
+      "[0.8352802991867065,],\n",
+      "[-0.940551221370697,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20137950778007507,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22083909809589386,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.5548670887947083,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07247430086135864,],\n",
+      "[0.4586115777492523,],\n",
+      "[-0.2480974942445755,],\n",
+      "[1.0230756998062134,],\n",
+      "[0.8352802991867065,],\n",
+      "[-0.940551221370697,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.29679017138551e-06,],\n",
+      "[-2.0575171220116317e-06,],\n",
+      "[-3.482636748231016e-05,],\n",
+      "[1.3960131582280155e-05,],\n",
+      "[-0.4341870844364166,],\n",
+      "[-0.010854240506887436,],\n",
+      "[-0.5979998111724854,],\n",
+      "[-0.2510700523853302,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20137950778007507,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9398701786994934,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07290849089622498,],\n",
+      "[0.45862242579460144,],\n",
+      "[-0.2474994957447052,],\n",
+      "[1.0233267545700073,],\n",
+      "[0.8353338837623596,],\n",
+      "[-0.9405622482299805,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20043963193893433,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1362123340368271,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07290849089622498,],\n",
+      "[0.45862242579460144,],\n",
+      "[-0.2474994957447052,],\n",
+      "[1.0233267545700073,],\n",
+      "[0.8353338837623596,],\n",
+      "[-0.9405622482299805,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.3198235819800175e-06,],\n",
+      "[-1.221463094225328e-06,],\n",
+      "[-2.1790716346004046e-05,],\n",
+      "[8.986270586319733e-06,],\n",
+      "[-0.4390532076358795,],\n",
+      "[-0.0110552329570055,],\n",
+      "[-0.6257294416427612,],\n",
+      "[-0.2501331865787506,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20043963193893433,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7381383776664734,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.0733475461602211,],\n",
+      "[0.4586334824562073,],\n",
+      "[-0.24687376618385315,],\n",
+      "[1.0235768556594849,],\n",
+      "[0.8353874683380127,],\n",
+      "[-0.9405732750892639,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1997014880180359,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0007464333903044462,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.0733475461602211,],\n",
+      "[0.4586334824562073,],\n",
+      "[-0.24687376618385315,],\n",
+      "[1.0235768556594849,],\n",
+      "[0.8353874683380127,],\n",
+      "[-0.9405732750892639,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.148701541836999e-08,],\n",
+      "[-6.831111676319779e-08,],\n",
+      "[-1.2849237691625603e-06,],\n",
+      "[5.452369009617541e-07,],\n",
+      "[-0.4514707922935486,],\n",
+      "[-0.01183705497533083,],\n",
+      "[-0.7233163118362427,],\n",
+      "[-0.24939803779125214,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1997014880180359,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.05464182049036026,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.0737990140914917,],\n",
+      "[0.4586453139781952,],\n",
+      "[-0.2461504489183426,],\n",
+      "[1.023826241493225,],\n",
+      "[0.8354410529136658,],\n",
+      "[-0.9405843019485474,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19964684545993805,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11946193873882294,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.0737990140914917,],\n",
+      "[0.4586453139781952,],\n",
+      "[-0.2461504489183426,],\n",
+      "[1.023826241493225,],\n",
+      "[0.8354410529136658,],\n",
+      "[-0.9405843019485474,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.617157737309753e-07,],\n",
+      "[6.516582402582571e-07,],\n",
+      "[1.2940920896653552e-05,],\n",
+      "[-5.653567313856911e-06,],\n",
+      "[-0.46330317854881287,],\n",
+      "[-0.012695159763097763,],\n",
+      "[-0.8293249607086182,],\n",
+      "[-0.24934488534927368,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19964684545993805,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6912649273872375,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07426232099533081,],\n",
+      "[0.45865800976753235,],\n",
+      "[-0.245321124792099,],\n",
+      "[1.0240756273269653,],\n",
+      "[0.8354946374893188,],\n",
+      "[-0.9405953288078308,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20033811032772064,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2717038691043854,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07426232099533081,],\n",
+      "[0.45865800976753235,],\n",
+      "[-0.245321124792099,],\n",
+      "[1.0240756273269653,],\n",
+      "[0.8354946374893188,],\n",
+      "[-0.9405953288078308,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.301803179871058e-07,],\n",
+      "[7.39559538942558e-07,],\n",
+      "[1.5527148207183927e-05,],\n",
+      "[-6.988339464442106e-06,],\n",
+      "[-0.4678522050380707,],\n",
+      "[-0.013132110238075256,],\n",
+      "[-0.8754134178161621,],\n",
+      "[-0.25003764033317566,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20033811032772064,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0425044298171997,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07473017275333405,],\n",
+      "[0.4586711525917053,],\n",
+      "[-0.24444571137428284,],\n",
+      "[1.0243256092071533,],\n",
+      "[0.8355482220649719,],\n",
+      "[-0.9406063556671143,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20138061046600342,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15297795832157135,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07473017275333405,],\n",
+      "[0.4586711525917053,],\n",
+      "[-0.24444571137428284,],\n",
+      "[1.0243256092071533,],\n",
+      "[0.8355482220649719,],\n",
+      "[-0.9406063556671143,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-4.008853125014866e-07,],\n",
+      "[4.1719295040820725e-07,],\n",
+      "[9.266604138247203e-06,],\n",
+      "[-4.297907707950799e-06,],\n",
+      "[-0.46333372592926025,],\n",
+      "[-0.012920585460960865,],\n",
+      "[-0.8360316157341003,],\n",
+      "[-0.25107714533805847,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20138061046600342,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7822472453117371,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07519350945949554,],\n",
+      "[0.4586840867996216,],\n",
+      "[-0.24360968172550201,],\n",
+      "[1.0245766639709473,],\n",
+      "[0.835601806640625,],\n",
+      "[-0.9406173825263977,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20216286182403564,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0010424922220408916,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07519350945949554,],\n",
+      "[0.4586840867996216,],\n",
+      "[-0.24360968172550201,],\n",
+      "[1.0245766639709473,],\n",
+      "[0.835601806640625,],\n",
+      "[-0.9406173825263977,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.4213999694211452e-08,],\n",
+      "[2.5909811185442777e-08,],\n",
+      "[6.085548420742271e-07,],\n",
+      "[-2.9079959062983107e-07,],\n",
+      "[-0.4533211290836334,],\n",
+      "[-0.012224124744534492,],\n",
+      "[-0.7412669658660889,],\n",
+      "[-0.2518593966960907,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20216286182403564,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.06457527726888657,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07564683258533478,],\n",
+      "[0.45869630575180054,],\n",
+      "[-0.24286840856075287,],\n",
+      "[1.024828553199768,],\n",
+      "[0.8356553912162781,],\n",
+      "[-0.9406284093856812,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20222744345664978,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11544226855039597,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07564683258533478,],\n",
+      "[0.45869630575180054,],\n",
+      "[-0.24286840856075287,],\n",
+      "[1.024828553199768,],\n",
+      "[0.8356553912162781,],\n",
+      "[-0.9406284093856812,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.8643919474925497e-07,],\n",
+      "[-2.0553028434733278e-07,],\n",
+      "[-5.097626399219735e-06,],\n",
+      "[2.5081571948248893e-06,],\n",
+      "[-0.44419458508491516,],\n",
+      "[-0.011491646990180016,],\n",
+      "[-0.6479165554046631,],\n",
+      "[-0.25192397832870483,],\n",
+      "[-0.05359897017478943,],\n",
+      "[0.011023559607565403,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20222744345664978,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6795352101325989,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07609102874994278,],\n",
+      "[0.4587078094482422,],\n",
+      "[-0.24222049117088318,],\n",
+      "[1.0250804424285889,],\n",
+      "[0.8357089757919312,],\n",
+      "[-0.9406394362449646,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20154790580272675,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [9/10], Loss: 0.1154\n",
+      "loss_antes tensor([0.49423569440841675,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6378946304321289,],\n",
+      "[-0.12578056752681732,],\n",
+      "[0.554867148399353,],\n",
+      "[-1.0009452104568481,],\n",
+      "[0.07609102874994278,],\n",
+      "[0.4587078094482422,],\n",
+      "[-0.24222049117088318,],\n",
+      "[1.0250804424285889,],\n",
+      "[0.8357089757919312,],\n",
+      "[-0.9406394362449646,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.28026294708251953,],\n",
+      "[-0.12703701853752136,],\n",
+      "[-0.3820388913154602,],\n",
+      "[0.05790916830301285,],\n",
+      "[-0.2936471998691559,],\n",
+      "[-0.008064738474786282,],\n",
+      "[-0.4787823259830475,],\n",
+      "[-0.1825977861881256,],\n",
+      "[-0.037699803709983826,],\n",
+      "[0.007893570698797703,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20154790580272675,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.40603768825531,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.637614369392395,],\n",
+      "[-0.12565353512763977,],\n",
+      "[0.5552492141723633,],\n",
+      "[-1.0010031461715698,],\n",
+      "[0.07638467848300934,],\n",
+      "[0.45871588587760925,],\n",
+      "[-0.24174170196056366,],\n",
+      "[1.0252630710601807,],\n",
+      "[0.8357466459274292,],\n",
+      "[-0.940647304058075,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.20014186203479767,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.48320645093917847,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.637614369392395,],\n",
+      "[-0.12565353512763977,],\n",
+      "[0.5552492141723633,],\n",
+      "[-1.0010031461715698,],\n",
+      "[0.07638467848300934,],\n",
+      "[0.45871588587760925,],\n",
+      "[-0.24174170196056366,],\n",
+      "[1.0252630710601807,],\n",
+      "[0.8357466459274292,],\n",
+      "[-0.940647304058075,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.29800862073898315,],\n",
+      "[-0.1255648136138916,],\n",
+      "[-0.3512035310268402,],\n",
+      "[0.05173952877521515,],\n",
+      "[-0.2947264611721039,],\n",
+      "[-0.007488698698580265,],\n",
+      "[-0.4803691506385803,],\n",
+      "[-0.19005367159843445,],\n",
+      "[-0.035464704036712646,],\n",
+      "[0.008774123154580593,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.20014186203479767,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.3902586698532104,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6373163461685181,],\n",
+      "[-0.12552796304225922,],\n",
+      "[0.5556004047393799,],\n",
+      "[-1.0010548830032349,],\n",
+      "[0.07667940855026245,],\n",
+      "[0.45872336626052856,],\n",
+      "[-0.2412613332271576,],\n",
+      "[1.0254530906677246,],\n",
+      "[0.8357821106910706,],\n",
+      "[-0.9406560659408569,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19875159859657288,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.8097985982894897,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6373163461685181,],\n",
+      "[-0.12552796304225922,],\n",
+      "[0.5556004047393799,],\n",
+      "[-1.0010548830032349,],\n",
+      "[0.07667940855026245,],\n",
+      "[0.45872336626052856,],\n",
+      "[-0.2412613332271576,],\n",
+      "[1.0254530906677246,],\n",
+      "[0.8357821106910706,],\n",
+      "[-0.9406560659408569,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.3957030773162842,],\n",
+      "[-0.15622767806053162,],\n",
+      "[-0.41332611441612244,],\n",
+      "[0.05969219282269478,],\n",
+      "[-0.24833416938781738,],\n",
+      "[-0.005863454658538103,],\n",
+      "[-0.4049674868583679,],\n",
+      "[-0.18134021759033203,],\n",
+      "[-0.028657205402851105,],\n",
+      "[0.00894987490028143,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19875159859657288,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.799774408340454,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6369206309318542,],\n",
+      "[-0.12537173926830292,],\n",
+      "[0.5560137033462524,],\n",
+      "[-1.0011146068572998,],\n",
+      "[0.07692774385213852,],\n",
+      "[0.45872923731803894,],\n",
+      "[-0.2408563643693924,],\n",
+      "[1.0256344079971313,],\n",
+      "[0.8358107805252075,],\n",
+      "[-0.9406650066375732,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19695182144641876,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.2104660272598267,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6369206309318542,],\n",
+      "[-0.12537173926830292,],\n",
+      "[0.5560137033462524,],\n",
+      "[-1.0011146068572998,],\n",
+      "[0.07692774385213852,],\n",
+      "[0.45872923731803894,],\n",
+      "[-0.2408563643693924,],\n",
+      "[1.0256344079971313,],\n",
+      "[0.8358107805252075,],\n",
+      "[-0.9406650066375732,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.4725396931171417,],\n",
+      "[-0.17689456045627594,],\n",
+      "[-0.450158953666687,],\n",
+      "[0.0642135813832283,],\n",
+      "[-0.20376312732696533,],\n",
+      "[-0.004315975122153759,],\n",
+      "[-0.33096203207969666,],\n",
+      "[-0.1767393797636032,],\n",
+      "[-0.022986125200986862,],\n",
+      "[0.009371201507747173,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19695182144641876,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.2004237174987793,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.636448085308075,],\n",
+      "[-0.12519484758377075,],\n",
+      "[0.5564638376235962,],\n",
+      "[-1.0011788606643677,],\n",
+      "[0.07713150978088379,],\n",
+      "[0.45873355865478516,],\n",
+      "[-0.24052540957927704,],\n",
+      "[1.0258111953735352,],\n",
+      "[0.8358337879180908,],\n",
+      "[-0.940674364566803,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19475139677524567,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([1.1787031888961792,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.636448085308075,],\n",
+      "[-0.12519484758377075,],\n",
+      "[0.5564638376235962,],\n",
+      "[-1.0011788606643677,],\n",
+      "[0.07713150978088379,],\n",
+      "[0.45873355865478516,],\n",
+      "[-0.24052540957927704,],\n",
+      "[1.0258111953735352,],\n",
+      "[0.8358337879180908,],\n",
+      "[-0.940674364566803,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.43409463763237,],\n",
+      "[-0.15622881054878235,],\n",
+      "[-0.3885321021080017,],\n",
+      "[0.0551312193274498,],\n",
+      "[-0.20811697840690613,],\n",
+      "[-0.0037608006969094276,],\n",
+      "[-0.3349820375442505,],\n",
+      "[-0.1867997944355011,],\n",
+      "[-0.02418985590338707,],\n",
+      "[0.010293788276612759,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19475139677524567,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-2.1713624000549316,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6360139846801758,],\n",
+      "[-0.12503862380981445,],\n",
+      "[0.5568523406982422,],\n",
+      "[-1.0012339353561401,],\n",
+      "[0.07733962684869766,],\n",
+      "[0.458737313747406,],\n",
+      "[-0.24019043147563934,],\n",
+      "[1.0259979963302612,],\n",
+      "[0.8358579874038696,],\n",
+      "[-0.9406846761703491,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19258002936840057,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.6119499802589417,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6360139846801758,],\n",
+      "[-0.12503862380981445,],\n",
+      "[0.5568523406982422,],\n",
+      "[-1.0012339353561401,],\n",
+      "[0.07733962684869766,],\n",
+      "[0.458737313747406,],\n",
+      "[-0.24019043147563934,],\n",
+      "[1.0259979963302612,],\n",
+      "[0.8358579874038696,],\n",
+      "[-0.9406846761703491,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.27840426564216614,],\n",
+      "[-0.09765033423900604,],\n",
+      "[-0.2408892810344696,],\n",
+      "[0.03423691168427467,],\n",
+      "[-0.276464581489563,],\n",
+      "[-0.004210997838526964,],\n",
+      "[-0.4441649913787842,],\n",
+      "[-0.20847390592098236,],\n",
+      "[-0.03253384307026863,],\n",
+      "[0.011385412886738777,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19258002936840057,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-1.5645467042922974,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357355713844299,],\n",
+      "[-0.12494097650051117,],\n",
+      "[0.5570932030677795,],\n",
+      "[-1.0012681484222412,],\n",
+      "[0.07761608809232712,],\n",
+      "[0.45874151587486267,],\n",
+      "[-0.23974627256393433,],\n",
+      "[1.0262064933776855,],\n",
+      "[0.835890531539917,],\n",
+      "[-0.9406960606575012,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19101548194885254,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.089415542781353,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357355713844299,],\n",
+      "[-0.12494097650051117,],\n",
+      "[0.5570932030677795,],\n",
+      "[-1.0012681484222412,],\n",
+      "[0.07761608809232712,],\n",
+      "[0.45874151587486267,],\n",
+      "[-0.23974627256393433,],\n",
+      "[1.0262064933776855,],\n",
+      "[0.835890531539917,],\n",
+      "[-0.9406960606575012,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.09102971106767654,],\n",
+      "[-0.031480204313993454,],\n",
+      "[-0.07811285555362701,],\n",
+      "[0.011191891506314278,],\n",
+      "[-0.3827398717403412,],\n",
+      "[-0.005129966884851456,],\n",
+      "[-0.6190211176872253,],\n",
+      "[-0.2313006967306137,],\n",
+      "[-0.04307030141353607,],\n",
+      "[0.012216857634484768,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19101548194885254,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.598048985004425,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356445550918579,],\n",
+      "[-0.12490949779748917,],\n",
+      "[0.5571713447570801,],\n",
+      "[-1.001279354095459,],\n",
+      "[0.0779988244175911,],\n",
+      "[0.45874664187431335,],\n",
+      "[-0.2391272485256195,],\n",
+      "[1.026437759399414,],\n",
+      "[0.8359336256980896,],\n",
+      "[-0.9407082796096802,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19041743874549866,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.018777839839458466,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356445550918579,],\n",
+      "[-0.12490949779748917,],\n",
+      "[0.5571713447570801,],\n",
+      "[-1.001279354095459,],\n",
+      "[0.0779988244175911,],\n",
+      "[0.45874664187431335,],\n",
+      "[-0.2391272485256195,],\n",
+      "[1.026437759399414,],\n",
+      "[0.8359336256980896,],\n",
+      "[-0.9407082796096802,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.03450434282422066,],\n",
+      "[0.011870048940181732,],\n",
+      "[0.03000614047050476,],\n",
+      "[-0.004357792902737856,],\n",
+      "[-0.4764968454837799,],\n",
+      "[-0.006100737024098635,],\n",
+      "[-0.7772327661514282,],\n",
+      "[-0.24618175625801086,],\n",
+      "[-0.05031475052237511,],\n",
+      "[0.012610564939677715,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19041743874549866,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.2740648090839386,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356790661811829,],\n",
+      "[-0.12492136657238007,],\n",
+      "[0.5571413636207581,],\n",
+      "[-1.0012749433517456,],\n",
+      "[0.07847531884908676,],\n",
+      "[0.45875275135040283,],\n",
+      "[-0.2383500188589096,],\n",
+      "[1.0266839265823364,],\n",
+      "[0.83598393201828,],\n",
+      "[-0.9407209157943726,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1906915009021759,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10126962512731552,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356790661811829,],\n",
+      "[-0.12492136657238007,],\n",
+      "[0.5571413636207581,],\n",
+      "[-1.0012749433517456,],\n",
+      "[0.07847531884908676,],\n",
+      "[0.45875275135040283,],\n",
+      "[-0.2383500188589096,],\n",
+      "[1.0266839265823364,],\n",
+      "[0.83598393201828,],\n",
+      "[-0.9407209157943726,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.06448245048522949,],\n",
+      "[0.022221000865101814,],\n",
+      "[0.057844895869493484,],\n",
+      "[-0.008550743572413921,],\n",
+      "[-0.514152467250824,],\n",
+      "[-0.006782816257327795,],\n",
+      "[-0.8433018326759338,],\n",
+      "[-0.2502651810646057,],\n",
+      "[-0.05213400349020958,],\n",
+      "[0.012659118510782719,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1906915009021759,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6364579200744629,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357435584068298,],\n",
+      "[-0.12494358420372009,],\n",
+      "[0.5570835471153259,],\n",
+      "[-1.001266360282898,],\n",
+      "[0.07898946851491928,],\n",
+      "[0.4587595462799072,],\n",
+      "[-0.2375067174434662,],\n",
+      "[1.0269341468811035,],\n",
+      "[0.8360360860824585,],\n",
+      "[-0.9407335519790649,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19132795929908752,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.03329324349761009,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357435584068298,],\n",
+      "[-0.12494358420372009,],\n",
+      "[0.5570835471153259,],\n",
+      "[-1.001266360282898,],\n",
+      "[0.07898946851491928,],\n",
+      "[0.4587595462799072,],\n",
+      "[-0.2375067174434662,],\n",
+      "[1.0269341468811035,],\n",
+      "[0.8360360860824585,],\n",
+      "[-0.9407335519790649,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.029097670689225197,],\n",
+      "[0.010108825750648975,],\n",
+      "[0.027314012870192528,],\n",
+      "[-0.004121961537748575,],\n",
+      "[-0.48519429564476013,],\n",
+      "[-0.006735005881637335,],\n",
+      "[-0.7942541837692261,],\n",
+      "[-0.24658532440662384,],\n",
+      "[-0.05013453960418701,],\n",
+      "[0.012559006921947002,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19132795929908752,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.364928662776947,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357726454734802,],\n",
+      "[-0.12495369464159012,],\n",
+      "[0.5570562481880188,],\n",
+      "[-1.0012621879577637,],\n",
+      "[0.07947466522455215,],\n",
+      "[0.45876628160476685,],\n",
+      "[-0.23671245574951172,],\n",
+      "[1.027180790901184,],\n",
+      "[0.8360862135887146,],\n",
+      "[-0.9407461285591125,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19169288873672485,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.02203705720603466,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357726454734802,],\n",
+      "[-0.12495369464159012,],\n",
+      "[0.5570562481880188,],\n",
+      "[-1.0012621879577637,],\n",
+      "[0.07947466522455215,],\n",
+      "[0.45876628160476685,],\n",
+      "[-0.23671245574951172,],\n",
+      "[1.027180790901184,],\n",
+      "[0.8360862135887146,],\n",
+      "[-0.9407461285591125,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.018296219408512115,],\n",
+      "[-0.006450865417718887,],\n",
+      "[-0.018179794773459435,],\n",
+      "[0.0028069133404642344,],\n",
+      "[-0.41779622435569763,],\n",
+      "[-0.005880517419427633,],\n",
+      "[-0.6740432977676392,],\n",
+      "[-0.240662083029747,],\n",
+      "[-0.0473969392478466,],\n",
+      "[0.012463490478694439,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19169288873672485,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.29689788818359375,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357543468475342,],\n",
+      "[-0.12494724243879318,],\n",
+      "[0.5570744276046753,],\n",
+      "[-1.001265048980713,],\n",
+      "[0.07989246398210526,],\n",
+      "[0.4587721526622772,],\n",
+      "[-0.2360384166240692,],\n",
+      "[1.027421474456787,],\n",
+      "[0.836133599281311,],\n",
+      "[-0.9407585859298706,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19139599800109863,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.19513703882694244,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357543468475342,],\n",
+      "[-0.12494724243879318,],\n",
+      "[0.5570744276046753,],\n",
+      "[-1.001265048980713,],\n",
+      "[0.07989246398210526,],\n",
+      "[0.4587721526622772,],\n",
+      "[-0.2360384166240692,],\n",
+      "[1.027421474456787,],\n",
+      "[0.836133599281311,],\n",
+      "[-0.9407585859298706,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.04147033020853996,],\n",
+      "[-0.014936438761651516,],\n",
+      "[-0.04401196166872978,],\n",
+      "[0.0069640446454286575,],\n",
+      "[-0.36069756746292114,],\n",
+      "[-0.0048460825346410275,],\n",
+      "[-0.5673952698707581,],\n",
+      "[-0.23686958849430084,],\n",
+      "[-0.04604379087686539,],\n",
+      "[0.012431764043867588,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19139599800109863,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.8834854960441589,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357128620147705,],\n",
+      "[-0.12493230402469635,],\n",
+      "[0.5571184158325195,],\n",
+      "[-1.0012719631195068,],\n",
+      "[0.08025316148996353,],\n",
+      "[0.4587770104408264,],\n",
+      "[-0.23547102510929108,],\n",
+      "[1.0276583433151245,],\n",
+      "[0.8361796140670776,],\n",
+      "[-0.9407710433006287,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1905125081539154,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.24392879009246826,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357128620147705,],\n",
+      "[-0.12493230402469635,],\n",
+      "[0.5571184158325195,],\n",
+      "[-1.0012719631195068,],\n",
+      "[0.08025316148996353,],\n",
+      "[0.4587770104408264,],\n",
+      "[-0.23547102510929108,],\n",
+      "[1.0276583433151245,],\n",
+      "[0.8361796140670776,],\n",
+      "[-0.9407710433006287,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.03492435812950134,],\n",
+      "[-0.012915701605379581,],\n",
+      "[-0.03984992206096649,],\n",
+      "[0.006470918655395508,],\n",
+      "[-0.35280969738960266,],\n",
+      "[-0.004384497180581093,],\n",
+      "[-0.5483208894729614,],\n",
+      "[-0.2366195172071457,],\n",
+      "[-0.04639201983809471,],\n",
+      "[0.012452531605958939,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1905125081539154,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9877815246582031,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356779336929321,],\n",
+      "[-0.12491938471794128,],\n",
+      "[0.5571582913398743,],\n",
+      "[-1.0012784004211426,],\n",
+      "[0.08060596883296967,],\n",
+      "[0.4587813913822174,],\n",
+      "[-0.23492270708084106,],\n",
+      "[1.0278949737548828,],\n",
+      "[0.8362259864807129,],\n",
+      "[-0.9407835006713867,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.18952472507953644,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0640627071261406,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356779336929321,],\n",
+      "[-0.12491938471794128,],\n",
+      "[0.5571582913398743,],\n",
+      "[-1.0012784004211426,],\n",
+      "[0.08060596883296967,],\n",
+      "[0.4587813913822174,],\n",
+      "[-0.23492270708084106,],\n",
+      "[1.0278949737548828,],\n",
+      "[0.8362259864807129,],\n",
+      "[-0.9407835006713867,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.013372959569096565,],\n",
+      "[-0.005091464146971703,],\n",
+      "[-0.01647799089550972,],\n",
+      "[0.002749056788161397,],\n",
+      "[-0.40004703402519226,],\n",
+      "[-0.00469267787411809,],\n",
+      "[-0.6353237628936768,],\n",
+      "[-0.2388344556093216,],\n",
+      "[-0.04758940264582634,],\n",
+      "[0.012490208260715008,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.18952472507953644,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.5062126517295837,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356645822525024,],\n",
+      "[-0.12491429597139359,],\n",
+      "[0.5571747422218323,],\n",
+      "[-1.0012811422348022,],\n",
+      "[0.0810060128569603,],\n",
+      "[0.4587860703468323,],\n",
+      "[-0.23428738117218018,],\n",
+      "[1.028133749961853,],\n",
+      "[0.8362735509872437,],\n",
+      "[-0.9407960176467896,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.18901851773262024,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.022544410079717636,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356645822525024,],\n",
+      "[-0.12491429597139359,],\n",
+      "[0.5571747422218323,],\n",
+      "[-1.0012811422348022,],\n",
+      "[0.0810060128569603,],\n",
+      "[0.4587860703468323,],\n",
+      "[-0.23428738117218018,],\n",
+      "[1.028133749961853,],\n",
+      "[0.8362735509872437,],\n",
+      "[-0.9407960176467896,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005894716363400221,],\n",
+      "[0.0023106832522898912,],\n",
+      "[0.007864643819630146,],\n",
+      "[-0.0013493150472640991,],\n",
+      "[-0.47388532757759094,],\n",
+      "[-0.005515489727258682,],\n",
+      "[-0.7808527946472168,],\n",
+      "[-0.2416069209575653,],\n",
+      "[-0.048657696694135666,],\n",
+      "[0.012517279013991356,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.18901851773262024,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.3002961277961731,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356704831123352,],\n",
+      "[-0.12491660565137863,],\n",
+      "[0.5571668744087219,],\n",
+      "[-1.0012798309326172,],\n",
+      "[0.08147989958524704,],\n",
+      "[0.458791583776474,],\n",
+      "[-0.23350653052330017,],\n",
+      "[1.0283753871917725,],\n",
+      "[0.8363221883773804,],\n",
+      "[-0.9408085346221924,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.18931882083415985,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2334422618150711,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356704831123352,],\n",
+      "[-0.12491660565137863,],\n",
+      "[0.5571668744087219,],\n",
+      "[-1.0012798309326172,],\n",
+      "[0.08147989958524704,],\n",
+      "[0.458791583776474,],\n",
+      "[-0.23350653052330017,],\n",
+      "[1.0283753871917725,],\n",
+      "[0.8363221883773804,],\n",
+      "[-0.9408085346221924,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.01404031552374363,],\n",
+      "[0.00565871549770236,],\n",
+      "[0.020314564928412437,],\n",
+      "[-0.0035870836582034826,],\n",
+      "[-0.5310424566268921,],\n",
+      "[-0.006493858061730862,],\n",
+      "[-0.9008142352104187,],\n",
+      "[-0.24357318878173828,],\n",
+      "[-0.049109090119600296,],\n",
+      "[0.01252562552690506,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.18931882083415985,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9663165807723999,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356845498085022,],\n",
+      "[-0.1249222606420517,],\n",
+      "[0.5571465492248535,],\n",
+      "[-1.0012762546539307,],\n",
+      "[0.08201093971729279,],\n",
+      "[0.4587980806827545,],\n",
+      "[-0.2326057106256485,],\n",
+      "[1.0286189317703247,],\n",
+      "[0.8363713026046753,],\n",
+      "[-0.9408210515975952,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19028513133525848,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.305746465921402,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356845498085022,],\n",
+      "[-0.1249222606420517,],\n",
+      "[0.5571465492248535,],\n",
+      "[-1.0012762546539307,],\n",
+      "[0.08201093971729279,],\n",
+      "[0.4587980806827545,],\n",
+      "[-0.2326057106256485,],\n",
+      "[1.0286189317703247,],\n",
+      "[0.8363713026046753,],\n",
+      "[-0.9408210515975952,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0118589261546731,],\n",
+      "[0.004908023867756128,],\n",
+      "[0.01862664334475994,],\n",
+      "[-0.003387180622667074,],\n",
+      "[-0.540298342704773,],\n",
+      "[-0.007152993232011795,],\n",
+      "[-0.9258823394775391,],\n",
+      "[-0.24431049823760986,],\n",
+      "[-0.048996467143297195,],\n",
+      "[0.012520774267613888,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19028513133525848,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.1058874130249023,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356964111328125,],\n",
+      "[-0.12492717057466507,],\n",
+      "[0.5571278929710388,],\n",
+      "[-1.0012729167938232,],\n",
+      "[0.08255124092102051,],\n",
+      "[0.45880523324012756,],\n",
+      "[-0.23167982697486877,],\n",
+      "[1.0288631916046143,],\n",
+      "[0.8364202976226807,],\n",
+      "[-0.940833568572998,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1913910210132599,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10500689595937729,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356964111328125,],\n",
+      "[-0.12492717057466507,],\n",
+      "[0.5571278929710388,],\n",
+      "[-1.0012729167938232,],\n",
+      "[0.08255124092102051,],\n",
+      "[0.45880523324012756,],\n",
+      "[-0.23167982697486877,],\n",
+      "[1.0288631916046143,],\n",
+      "[0.8364202976226807,],\n",
+      "[-0.940833568572998,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.005116885993629694,],\n",
+      "[0.0021748575381934643,],\n",
+      "[0.008732368238270283,],\n",
+      "[-0.001635948894545436,],\n",
+      "[-0.5002043843269348,],\n",
+      "[-0.007025589235126972,],\n",
+      "[-0.8438615202903748,],\n",
+      "[-0.24407972395420074,],\n",
+      "[-0.048640042543411255,],\n",
+      "[0.012512107379734516,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1913910210132599,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6480950713157654,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357015371322632,],\n",
+      "[-0.12492934614419937,],\n",
+      "[0.5571191310882568,],\n",
+      "[-1.0012712478637695,],\n",
+      "[0.08305144309997559,],\n",
+      "[0.45881226658821106,],\n",
+      "[-0.2308359593153,],\n",
+      "[1.0291072130203247,],\n",
+      "[0.8364689350128174,],\n",
+      "[-0.9408460855484009,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1920391172170639,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.003526984713971615,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357015371322632,],\n",
+      "[-0.12492934614419937,],\n",
+      "[0.5571191310882568,],\n",
+      "[-1.0012712478637695,],\n",
+      "[0.08305144309997559,],\n",
+      "[0.45881226658821106,],\n",
+      "[-0.2308359593153,],\n",
+      "[1.0291072130203247,],\n",
+      "[0.8364689350128174,],\n",
+      "[-0.9408460855484009,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0006890507647767663,],\n",
+      "[-0.0003012801753357053,],\n",
+      "[-0.0012787748128175735,],\n",
+      "[0.00024683246738277376,],\n",
+      "[-0.4383583068847656,],\n",
+      "[-0.006166051607578993,],\n",
+      "[-0.7071672081947327,],\n",
+      "[-0.24336989223957062,],\n",
+      "[-0.04833470284938812,],\n",
+      "[0.012505825608968735,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1920391172170639,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.11877696216106415,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6357008218765259,],\n",
+      "[-0.1249290481209755,],\n",
+      "[0.5571203827857971,],\n",
+      "[-1.0012714862823486,],\n",
+      "[0.0834897980093956,],\n",
+      "[0.4588184356689453,],\n",
+      "[-0.23012879490852356,],\n",
+      "[1.0293506383895874,],\n",
+      "[0.8365172743797302,],\n",
+      "[-0.9408586025238037,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19192034006118774,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13012121617794037,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6357008218765259,],\n",
+      "[-0.1249290481209755,],\n",
+      "[0.5571203827857971,],\n",
+      "[-1.0012714862823486,],\n",
+      "[0.0834897980093956,],\n",
+      "[0.4588184356689453,],\n",
+      "[-0.23012879490852356,],\n",
+      "[1.0293506383895874,],\n",
+      "[0.8365172743797302,],\n",
+      "[-0.9408586025238037,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.003070330712944269,],\n",
+      "[-0.0013843729393556714,],\n",
+      "[-0.006201429758220911,],\n",
+      "[0.0012331366306170821,],\n",
+      "[-0.39338359236717224,],\n",
+      "[-0.005257569719105959,],\n",
+      "[-0.6003828048706055,],\n",
+      "[-0.242582768201828,],\n",
+      "[-0.0482102632522583,],\n",
+      "[0.01250382885336876,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19192034006118774,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.721446692943573,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492766231298447,],\n",
+      "[0.5571265816688538,],\n",
+      "[-1.0012726783752441,],\n",
+      "[0.08388318121433258,],\n",
+      "[0.45882368087768555,],\n",
+      "[-0.22952841222286224,],\n",
+      "[1.0295932292938232,],\n",
+      "[0.8365654945373535,],\n",
+      "[-0.9408711194992065,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1911989003419876,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15540103614330292,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492766231298447,],\n",
+      "[0.5571265816688538,],\n",
+      "[-1.0012726783752441,],\n",
+      "[0.08388318121433258,],\n",
+      "[0.45882368087768555,],\n",
+      "[-0.22952841222286224,],\n",
+      "[1.0295932292938232,],\n",
+      "[0.8365654945373535,],\n",
+      "[-0.9408711194992065,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.002458745613694191,],\n",
+      "[-0.0011453218758106232,],\n",
+      "[-0.005407804623246193,],\n",
+      "[0.0011075909715145826,],\n",
+      "[-0.3907886743545532,],\n",
+      "[-0.004969581961631775,],\n",
+      "[-0.5886921286582947,],\n",
+      "[-0.24197357892990112,],\n",
+      "[-0.04824046418070793,],\n",
+      "[0.012504697777330875,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1911989003419876,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7884188890457153,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356952786445618,],\n",
+      "[-0.12492651492357254,],\n",
+      "[0.5571320056915283,],\n",
+      "[-1.00127375125885,],\n",
+      "[0.08427397161722183,],\n",
+      "[0.4588286578655243,],\n",
+      "[-0.2289397269487381,],\n",
+      "[1.0298352241516113,],\n",
+      "[0.8366137146949768,],\n",
+      "[-0.9408836364746094,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19041047990322113,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.018930919468402863,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356952786445618,],\n",
+      "[-0.12492651492357254,],\n",
+      "[0.5571320056915283,],\n",
+      "[-1.00127375125885,],\n",
+      "[0.08427397161722183,],\n",
+      "[0.4588286578655243,],\n",
+      "[-0.2289397269487381,],\n",
+      "[1.0298352241516113,],\n",
+      "[0.8366137146949768,],\n",
+      "[-0.9408836364746094,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.0006284178234636784,],\n",
+      "[-0.0003024909528903663,],\n",
+      "[-0.0015052647795528173,],\n",
+      "[0.000317555881338194,],\n",
+      "[-0.4284462332725525,],\n",
+      "[-0.005420921370387077,],\n",
+      "[-0.679007351398468,],\n",
+      "[-0.24171718955039978,],\n",
+      "[-0.04833246022462845,],\n",
+      "[0.012506443075835705,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19041047990322113,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.2751796543598175,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356946229934692,],\n",
+      "[-0.12492620944976807,],\n",
+      "[0.5571334958076477,],\n",
+      "[-1.0012741088867188,],\n",
+      "[0.08470241725444794,],\n",
+      "[0.45883408188819885,],\n",
+      "[-0.22826072573661804,],\n",
+      "[1.0300769805908203,],\n",
+      "[0.8366620540618896,],\n",
+      "[-0.9408961534500122,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19013530015945435,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.06476940959692001,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356946229934692,],\n",
+      "[-0.12492620944976807,],\n",
+      "[0.5571334958076477,],\n",
+      "[-1.0012741088867188,],\n",
+      "[0.08470241725444794,],\n",
+      "[0.45883408188819885,],\n",
+      "[-0.22826072573661804,],\n",
+      "[1.0300769805908203,],\n",
+      "[0.8366620540618896,],\n",
+      "[-0.9408961534500122,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0008508404134772718,],\n",
+      "[0.00042262187344022095,],\n",
+      "[0.002218948444351554,],\n",
+      "[-0.0004823226772714406,],\n",
+      "[-0.48058104515075684,],\n",
+      "[-0.006307727657258511,],\n",
+      "[-0.8161914944648743,],\n",
+      "[-0.2419465184211731,],\n",
+      "[-0.04840630292892456,],\n",
+      "[0.012507599778473377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19013530015945435,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5089969038963318,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356954574584961,],\n",
+      "[-0.12492663413286209,],\n",
+      "[0.557131290435791,],\n",
+      "[-1.0012736320495605,],\n",
+      "[0.08518299460411072,],\n",
+      "[0.45884039998054504,],\n",
+      "[-0.22744452953338623,],\n",
+      "[1.0303189754486084,],\n",
+      "[0.8367104530334473,],\n",
+      "[-0.940908670425415,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1906442940235138,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.29648447036743164,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356954574584961,],\n",
+      "[-0.12492663413286209,],\n",
+      "[0.557131290435791,],\n",
+      "[-1.0012736320495605,],\n",
+      "[0.08518299460411072,],\n",
+      "[0.45884039998054504,],\n",
+      "[-0.22744452953338623,],\n",
+      "[1.0303189754486084,],\n",
+      "[0.8367104530334473,],\n",
+      "[-0.940908670425415,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0013321482110768557,],\n",
+      "[0.0006813306245021522,],\n",
+      "[0.0037806632462888956,],\n",
+      "[-0.0008471292676404119,],\n",
+      "[-0.5154443383216858,],\n",
+      "[-0.007192856166511774,],\n",
+      "[-0.9165660738945007,],\n",
+      "[-0.24266885221004486,],\n",
+      "[-0.048430655151605606,],\n",
+      "[0.012507829815149307,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1906442940235138,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0890074968338013,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356967687606812,],\n",
+      "[-0.12492731213569641,],\n",
+      "[0.5571275353431702,],\n",
+      "[-1.0012727975845337,],\n",
+      "[0.0856984406709671,],\n",
+      "[0.4588475823402405,],\n",
+      "[-0.22652795910835266,],\n",
+      "[1.0305616855621338,],\n",
+      "[0.8367589116096497,],\n",
+      "[-0.9409211874008179,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19173330068588257,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.3094882667064667,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356967687606812,],\n",
+      "[-0.12492731213569641,],\n",
+      "[0.5571275353431702,],\n",
+      "[-1.0012727975845337,],\n",
+      "[0.0856984406709671,],\n",
+      "[0.4588475823402405,],\n",
+      "[-0.22652795910835266,],\n",
+      "[1.0305616855621338,],\n",
+      "[0.8367589116096497,],\n",
+      "[-0.9409211874008179,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.0009957813890650868,],\n",
+      "[0.0005235772114247084,],\n",
+      "[0.0030741021037101746,],\n",
+      "[-0.0007103240350261331,],\n",
+      "[-0.5140751600265503,],\n",
+      "[-0.007601348217576742,],\n",
+      "[-0.9195035696029663,],\n",
+      "[-0.24366186559200287,],\n",
+      "[-0.04841450974345207,],\n",
+      "[0.012507477775216103,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19173330068588257,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.1126329898834229,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977820396423,],\n",
+      "[-0.1249278336763382,],\n",
+      "[0.5571244359016418,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.08621251583099365,],\n",
+      "[0.45885518193244934,],\n",
+      "[-0.2256084531545639,],\n",
+      "[1.0308053493499756,],\n",
+      "[0.8368073105812073,],\n",
+      "[-0.9409337043762207,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19284594058990479,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07971466332674026,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977820396423,],\n",
+      "[-0.1249278336763382,],\n",
+      "[0.5571244359016418,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.08621251583099365,],\n",
+      "[0.45885518193244934,],\n",
+      "[-0.2256084531545639,],\n",
+      "[1.0308053493499756,],\n",
+      "[0.8368073105812073,],\n",
+      "[-0.9409337043762207,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00036966250627301633,],\n",
+      "[0.0001998037041630596,],\n",
+      "[0.0012412676587700844,],\n",
+      "[-0.0002957880205940455,],\n",
+      "[-0.4799911677837372,],\n",
+      "[-0.007246742025017738,],\n",
+      "[-0.8248888850212097,],\n",
+      "[-0.24452391266822815,],\n",
+      "[-0.04838496819138527,],\n",
+      "[0.012506961822509766,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19284594058990479,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5646758675575256,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.635698139667511,],\n",
+      "[-0.12492803484201431,],\n",
+      "[0.5571231842041016,],\n",
+      "[-1.0012718439102173,],\n",
+      "[0.08669250458478928,],\n",
+      "[0.45886242389678955,],\n",
+      "[-0.22478356957435608,],\n",
+      "[1.0310498476028442,],\n",
+      "[0.8368557095527649,],\n",
+      "[-0.9409462213516235,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19341062009334564,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.012497939169406891,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.635698139667511,],\n",
+      "[-0.12492803484201431,],\n",
+      "[0.5571231842041016,],\n",
+      "[-1.0012718439102173,],\n",
+      "[0.08669250458478928,],\n",
+      "[0.45886242389678955,],\n",
+      "[-0.22478356957435608,],\n",
+      "[1.0310498476028442,],\n",
+      "[0.8368557095527649,],\n",
+      "[-0.9409462213516235,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00010704409214667976,],\n",
+      "[-5.95656456425786e-05,],\n",
+      "[-0.0003910798404831439,],\n",
+      "[9.607315587345511e-05,],\n",
+      "[-0.4356347620487213,],\n",
+      "[-0.006367464549839497,],\n",
+      "[-0.6915807127952576,],\n",
+      "[-0.2448628842830658,],\n",
+      "[-0.04836224019527435,],\n",
+      "[0.012506664730608463,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19341062009334564,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.22358855605125427,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356980204582214,],\n",
+      "[-0.12492797523736954,],\n",
+      "[0.557123601436615,],\n",
+      "[-1.0012719631195068,],\n",
+      "[0.08712814003229141,],\n",
+      "[0.4588688015937805,],\n",
+      "[-0.22409199178218842,],\n",
+      "[1.0312947034835815,],\n",
+      "[0.8369040489196777,],\n",
+      "[-0.9409587383270264,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19318702816963196,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15030767023563385,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356980204582214,],\n",
+      "[-0.12492797523736954,],\n",
+      "[0.557123601436615,],\n",
+      "[-1.0012719631195068,],\n",
+      "[0.08712814003229141,],\n",
+      "[0.4588688015937805,],\n",
+      "[-0.22409199178218842,],\n",
+      "[1.0312947034835815,],\n",
+      "[0.8369040489196777,],\n",
+      "[-0.9409587383270264,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00027143879560753703,],\n",
+      "[-0.00015583925414830446,],\n",
+      "[-0.0010795308044180274,],\n",
+      "[0.0002732441935222596,],\n",
+      "[-0.4077647626399994,],\n",
+      "[-0.005615234840661287,],\n",
+      "[-0.5995872020721436,],\n",
+      "[-0.2445441484451294,],\n",
+      "[-0.04835468530654907,],\n",
+      "[0.012506576254963875,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19318702816963196,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7753908038139343,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.124927818775177,],\n",
+      "[0.557124674320221,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.08753590285778046,],\n",
+      "[0.4588744044303894,],\n",
+      "[-0.22349239885807037,],\n",
+      "[1.0315392017364502,],\n",
+      "[0.8369523882865906,],\n",
+      "[-0.9409712553024292,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1924116313457489,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.14357973635196686,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.124927818775177,],\n",
+      "[0.557124674320221,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.08753590285778046,],\n",
+      "[0.4588744044303894,],\n",
+      "[-0.22349239885807037,],\n",
+      "[1.0315392017364502,],\n",
+      "[0.8369523882865906,],\n",
+      "[-0.9409712553024292,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[0.00019396527204662561,],\n",
+      "[-0.00011506692680995911,],\n",
+      "[-0.0008400400402024388,],\n",
+      "[0.00021898488921578974,],\n",
+      "[-0.41075220704078674,],\n",
+      "[-0.005490211304277182,],\n",
+      "[-0.6026694774627686,],\n",
+      "[-0.24380214512348175,],\n",
+      "[-0.04835814610123634,],\n",
+      "[0.012506680563092232,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1924116313457489,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7578377723693848,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356975436210632,],\n",
+      "[-0.12492770701646805,],\n",
+      "[0.5571255087852478,],\n",
+      "[-1.001272439956665,],\n",
+      "[0.08794665336608887,],\n",
+      "[0.45887988805770874,],\n",
+      "[-0.2228897362947464,],\n",
+      "[1.0317829847335815,],\n",
+      "[0.8370007276535034,],\n",
+      "[-0.940983772277832,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1916537880897522,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.008468644693493843,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356975436210632,],\n",
+      "[-0.12492770701646805,],\n",
+      "[0.5571255087852478,],\n",
+      "[-1.001272439956665,],\n",
+      "[0.08794665336608887,],\n",
+      "[0.45887988805770874,],\n",
+      "[-0.2228897362947464,],\n",
+      "[1.0317829847335815,],\n",
+      "[0.8370007276535034,],\n",
+      "[-0.940983772277832,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[3.44394393323455e-05,],\n",
+      "[-2.1109437511768192e-05,],\n",
+      "[-0.00016241525008808821,],\n",
+      "[4.3605057726381347e-05,],\n",
+      "[-0.4391252100467682,],\n",
+      "[-0.006001683417707682,],\n",
+      "[-0.6976469159126282,],\n",
+      "[-0.2431386560201645,],\n",
+      "[-0.048365410417318344,],\n",
+      "[0.01250678114593029,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1916537880897522,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.1840503215789795,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356974840164185,],\n",
+      "[-0.12492768466472626,],\n",
+      "[0.5571256875991821,],\n",
+      "[-1.001272439956665,],\n",
+      "[0.08838577568531036,],\n",
+      "[0.45888587832450867,],\n",
+      "[-0.22219209372997284,],\n",
+      "[1.0320261716842651,],\n",
+      "[0.8370490670204163,],\n",
+      "[-0.9409962892532349,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19146974384784698,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.08841003477573395,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356974840164185,],\n",
+      "[-0.12492768466472626,],\n",
+      "[0.5571256875991821,],\n",
+      "[-1.001272439956665,],\n",
+      "[0.08838577568531036,],\n",
+      "[0.45888587832450867,],\n",
+      "[-0.22219209372997284,],\n",
+      "[1.0320261716842651,],\n",
+      "[0.8370490670204163,],\n",
+      "[-0.9409962892532349,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.13510050647892e-05,],\n",
+      "[5.1439288654364645e-05,],\n",
+      "[0.00041756784776225686,],\n",
+      "[-0.00011550819908734411,],\n",
+      "[-0.4733496904373169,],\n",
+      "[-0.006814446300268173,],\n",
+      "[-0.8254683613777161,],\n",
+      "[-0.24303516745567322,],\n",
+      "[-0.048370592296123505,],\n",
+      "[0.012506782077252865,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19146974384784698,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.5946760773658752,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356975436210632,],\n",
+      "[-0.12492773681879044,],\n",
+      "[0.5571252703666687,],\n",
+      "[-1.0012723207473755,],\n",
+      "[0.08885912597179413,],\n",
+      "[0.45889270305633545,],\n",
+      "[-0.22136662900447845,],\n",
+      "[1.0322692394256592,],\n",
+      "[0.8370974659919739,],\n",
+      "[-0.9410088062286377,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19206441938877106,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.30370381474494934,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356975436210632,],\n",
+      "[-0.12492773681879044,],\n",
+      "[0.5571252703666687,],\n",
+      "[-1.0012723207473755,],\n",
+      "[0.08885912597179413,],\n",
+      "[0.45889270305633545,],\n",
+      "[-0.22136662900447845,],\n",
+      "[1.0322692394256592,],\n",
+      "[0.8370974659919739,],\n",
+      "[-0.9410088062286377,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-0.00011022856051567942,],\n",
+      "[7.17489529051818e-05,],\n",
+      "[0.0006154353031888604,],\n",
+      "[-0.00017550960183143616,],\n",
+      "[-0.4929080605506897,],\n",
+      "[-0.007494139950722456,],\n",
+      "[-0.9065877795219421,],\n",
+      "[-0.24365642666816711,],\n",
+      "[-0.04837178438901901,],\n",
+      "[0.012506779283285141,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19206441938877106,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.1021863222122192,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356976628303528,],\n",
+      "[-0.1249278113245964,],\n",
+      "[0.557124674320221,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.08935203403234482,],\n",
+      "[0.45890018343925476,],\n",
+      "[-0.22046004235744476,],\n",
+      "[1.032512903213501,],\n",
+      "[0.8371458649635315,],\n",
+      "[-0.9410213232040405,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1931665986776352,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2635645270347595,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356976628303528,],\n",
+      "[-0.1249278113245964,],\n",
+      "[0.557124674320221,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.08935203403234482,],\n",
+      "[0.45890018343925476,],\n",
+      "[-0.22046004235744476,],\n",
+      "[1.032512903213501,],\n",
+      "[0.8371458649635315,],\n",
+      "[-0.9410213232040405,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.506897964049131e-05,],\n",
+      "[5.023063567932695e-05,],\n",
+      "[0.00045570993097499013,],\n",
+      "[-0.000134032525238581,],\n",
+      "[-0.4877815246582031,],\n",
+      "[-0.007661859970539808,],\n",
+      "[-0.891933262348175,],\n",
+      "[-0.2447374016046524,],\n",
+      "[-0.048370394855737686,],\n",
+      "[0.012506775557994843,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1931665986776352,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0267690420150757,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571241974830627,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.08983981609344482,],\n",
+      "[0.4589078426361084,],\n",
+      "[-0.21956810355186462,],\n",
+      "[1.0327576398849487,],\n",
+      "[0.8371942639350891,],\n",
+      "[-0.9410338401794434,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19419336318969727,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.04212040454149246,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571241974830627,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.08983981609344482,],\n",
+      "[0.4589078426361084,],\n",
+      "[-0.21956810355186462,],\n",
+      "[1.0327576398849487,],\n",
+      "[0.8371942639350891,],\n",
+      "[-0.9410338401794434,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.193833643104881e-05,],\n",
+      "[1.5093173715285957e-05,],\n",
+      "[0.00014480415848083794,],\n",
+      "[-4.392192204250023e-05,],\n",
+      "[-0.46301645040512085,],\n",
+      "[-0.007207517046481371,],\n",
+      "[-0.7935270071029663,],\n",
+      "[-0.24572069942951202,],\n",
+      "[-0.04836822673678398,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19419336318969727,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.4104645848274231,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492787837982178,],\n",
+      "[0.5571240782737732,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.09030283242464066,],\n",
+      "[0.4589150547981262,],\n",
+      "[-0.21877457201480865,],\n",
+      "[1.033003330230713,],\n",
+      "[0.837242603302002,],\n",
+      "[-0.9410463571548462,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.194603830575943,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0354502834379673,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492787837982178,],\n",
+      "[0.5571240782737732,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.09030283242464066,],\n",
+      "[0.4589150547981262,],\n",
+      "[-0.21877457201480865,],\n",
+      "[1.033003330230713,],\n",
+      "[0.837242603302002,],\n",
+      "[-0.9410463571548462,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.4712914889969397e-05,],\n",
+      "[-1.0424928404972889e-05,],\n",
+      "[-0.00010563979594735429,],\n",
+      "[3.302843833807856e-05,],\n",
+      "[-0.4347117245197296,],\n",
+      "[-0.00644277036190033,],\n",
+      "[-0.6725576519966125,],\n",
+      "[-0.24609531462192535,],\n",
+      "[-0.04836638271808624,],\n",
+      "[0.012506772764027119,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.194603830575943,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.37656500935554504,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492787092924118,],\n",
+      "[0.5571241974830627,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.09073754400014877,],\n",
+      "[0.45892149209976196,],\n",
+      "[-0.21810200810432434,],\n",
+      "[1.0332493782043457,],\n",
+      "[0.8372909426689148,],\n",
+      "[-0.941058874130249,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19422726333141327,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.18515855073928833,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492787092924118,],\n",
+      "[0.5571241974830627,],\n",
+      "[-1.0012720823287964,],\n",
+      "[0.09073754400014877,],\n",
+      "[0.45892149209976196,],\n",
+      "[-0.21810200810432434,],\n",
+      "[1.0332493782043457,],\n",
+      "[0.8372909426689148,],\n",
+      "[-0.941058874130249,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.458020026097074e-05,],\n",
+      "[-1.7975904484046623e-05,],\n",
+      "[-0.0001921017246786505,],\n",
+      "[6.186863902257755e-05,],\n",
+      "[-0.41958174109458923,],\n",
+      "[-0.005895459093153477,],\n",
+      "[-0.5995577573776245,],\n",
+      "[-0.24570688605308533,],\n",
+      "[-0.0483659990131855,],\n",
+      "[0.012506772764027119,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19422726333141327,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.860601544380188,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492785602807999,],\n",
+      "[0.5571243762969971,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09115712344646454,],\n",
+      "[0.4589273929595947,],\n",
+      "[-0.2175024449825287,],\n",
+      "[1.0334950685501099,],\n",
+      "[0.8373392820358276,],\n",
+      "[-0.9410713911056519,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1933666616678238,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.1412641555070877,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492785602807999,],\n",
+      "[0.5571243762969971,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09115712344646454,],\n",
+      "[0.4589273929595947,],\n",
+      "[-0.2175024449825287,],\n",
+      "[1.0334950685501099,],\n",
+      "[0.8373392820358276,],\n",
+      "[-0.9410713911056519,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.569462074257899e-05,],\n",
+      "[-1.1861971870530397e-05,],\n",
+      "[-0.00013356101408135146,],\n",
+      "[4.4291653466643766e-05,],\n",
+      "[-0.42466601729393005,],\n",
+      "[-0.005891864188015461,],\n",
+      "[-0.6157881021499634,],\n",
+      "[-0.24485385417938232,],\n",
+      "[-0.048366349190473557,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1933666616678238,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7517035007476807,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492784112691879,],\n",
+      "[0.5571244955062866,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.091581791639328,],\n",
+      "[0.4589332938194275,],\n",
+      "[-0.21688665449619293,],\n",
+      "[1.0337399244308472,],\n",
+      "[0.8373876214027405,],\n",
+      "[-0.9410839080810547,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19261495769023895,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.003593983594328165,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492784112691879,],\n",
+      "[0.5571244955062866,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.091581791639328,],\n",
+      "[0.4589332938194275,],\n",
+      "[-0.21688665449619293,],\n",
+      "[1.0337399244308472,],\n",
+      "[0.8373876214027405,],\n",
+      "[-0.9410839080810547,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.8299624571227469e-06,],\n",
+      "[-1.4289524870036985e-06,],\n",
+      "[-1.695554419711698e-05,],\n",
+      "[5.790165232610889e-06,],\n",
+      "[-0.4442649185657501,],\n",
+      "[-0.006371753290295601,],\n",
+      "[-0.7106355428695679,],\n",
+      "[-0.2441186159849167,],\n",
+      "[-0.048367079347372055,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19261495769023895,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.11989980936050415,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492784112691879,],\n",
+      "[0.5571244955062866,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09202605485916138,],\n",
+      "[0.45893967151641846,],\n",
+      "[-0.21617601811885834,],\n",
+      "[1.0339840650558472,],\n",
+      "[0.8374359607696533,],\n",
+      "[-0.9410964250564575,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19249506294727325,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.10465536266565323,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492784112691879,],\n",
+      "[0.5571244955062866,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09202605485916138,],\n",
+      "[0.45893967151641846,],\n",
+      "[-0.21617601811885834,],\n",
+      "[1.0339840650558472,],\n",
+      "[0.8374359607696533,],\n",
+      "[-0.9410964250564575,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-7.218624887173064e-06,],\n",
+      "[5.813684310851386e-06,],\n",
+      "[7.278429984580725e-05,],\n",
+      "[-2.5608100258978084e-05,],\n",
+      "[-0.4651719033718109,],\n",
+      "[-0.00702304532751441,],\n",
+      "[-0.8247232437133789,],\n",
+      "[-0.24401213228702545,],\n",
+      "[-0.0483674518764019,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19249506294727325,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6470101475715637,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492784857749939,],\n",
+      "[0.5571244359016418,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09249122440814972,],\n",
+      "[0.45894670486450195,],\n",
+      "[-0.21535129845142365,],\n",
+      "[1.0342280864715576,],\n",
+      "[0.8374843001365662,],\n",
+      "[-0.9411089420318604,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19314207136631012,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2911151945590973,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492784857749939,],\n",
+      "[0.5571244359016418,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09249122440814972,],\n",
+      "[0.45894670486450195,],\n",
+      "[-0.21535129845142365,],\n",
+      "[1.0342280864715576,],\n",
+      "[0.8374843001365662,],\n",
+      "[-0.9411089420318604,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-8.800854629953392e-06,],\n",
+      "[7.295202976820292e-06,],\n",
+      "[9.65037033893168e-05,],\n",
+      "[-3.500423190416768e-05,],\n",
+      "[-0.47519439458847046,],\n",
+      "[-0.007482111919671297,],\n",
+      "[-0.8857676982879639,],\n",
+      "[-0.2446620762348175,],\n",
+      "[-0.0483674593269825,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19314207136631012,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0791019201278687,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492785602807999,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09296642243862152,],\n",
+      "[0.45895418524742126,],\n",
+      "[-0.21446552872657776,],\n",
+      "[1.0344727039337158,],\n",
+      "[0.837532639503479,],\n",
+      "[-0.9411214590072632,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19422116875648499,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.20700128376483917,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492785602807999,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09296642243862152,],\n",
+      "[0.45895418524742126,],\n",
+      "[-0.21446552872657776,],\n",
+      "[1.0344727039337158,],\n",
+      "[0.837532639503479,],\n",
+      "[-0.9411214590072632,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-5.424969003797742e-06,],\n",
+      "[4.622871074388968e-06,],\n",
+      "[6.466839113272727e-05,],\n",
+      "[-2.419134216324892e-05,],\n",
+      "[-0.46966275572776794,],\n",
+      "[-0.0074925716035068035,],\n",
+      "[-0.8578686714172363,],\n",
+      "[-0.24573665857315063,],\n",
+      "[-0.048367466777563095,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19422116875648499,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.9099482297897339,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571242570877075,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09343608468770981,],\n",
+      "[0.4589616656303406,],\n",
+      "[-0.21360765397548676,],\n",
+      "[1.03471839427948,],\n",
+      "[0.8375809788703918,],\n",
+      "[-0.941133975982666,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1951311230659485,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.014162937179207802,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571242570877075,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09343608468770981,],\n",
+      "[0.4589616656303406,],\n",
+      "[-0.21360765397548676,],\n",
+      "[1.03471839427948,],\n",
+      "[0.8375809788703918,],\n",
+      "[-0.941133975982666,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-1.0373017858000821e-06,],\n",
+      "[9.091139645533985e-07,],\n",
+      "[1.3444157048070338e-05,],\n",
+      "[-5.1859879022231326e-06,],\n",
+      "[-0.4532516896724701,],\n",
+      "[-0.007057691924273968,],\n",
+      "[-0.7629532814025879,],\n",
+      "[-0.24663911759853363,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.1951311230659485,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.238016277551651,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571242570877075,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09388933330774307,],\n",
+      "[0.45896872878074646,],\n",
+      "[-0.21284469962120056,],\n",
+      "[1.0349650382995605,],\n",
+      "[0.8376293182373047,],\n",
+      "[-0.9411464929580688,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19536913931369781,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.07114461064338684,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571242570877075,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09388933330774307,],\n",
+      "[0.45896872878074646,],\n",
+      "[-0.21284469962120056,],\n",
+      "[1.0349650382995605,],\n",
+      "[0.8376293182373047,],\n",
+      "[-0.9411464929580688,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.6994811176118674e-06,],\n",
+      "[-1.5347068256232888e-06,],\n",
+      "[-2.3961854822118767e-05,],\n",
+      "[9.525801942800172e-06,],\n",
+      "[-0.43653932213783264,],\n",
+      "[-0.006471780128777027,],\n",
+      "[-0.6594498753547668,],\n",
+      "[-0.2468726485967636,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19536913931369781,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.533458948135376,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571242570877075,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09432587027549744,],\n",
+      "[0.4589751958847046,],\n",
+      "[-0.21218524873256683,],\n",
+      "[1.0352119207382202,],\n",
+      "[0.8376776576042175,],\n",
+      "[-0.9411590099334717,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19483567774295807,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.22125262022018433,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571242570877075,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09432587027549744,],\n",
+      "[0.4589751958847046,],\n",
+      "[-0.21218524873256683,],\n",
+      "[1.0352119207382202,],\n",
+      "[0.8376776576042175,],\n",
+      "[-0.9411590099334717,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[2.1908133476244984e-06,],\n",
+      "[-2.0428033167263493e-06,],\n",
+      "[-3.362498682690784e-05,],\n",
+      "[1.3767196833214257e-05,],\n",
+      "[-0.4291693866252899,],\n",
+      "[-0.006115353666245937,],\n",
+      "[-0.6055496335029602,],\n",
+      "[-0.24633771181106567,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19483567774295807,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.9407508373260498,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09475503861904144,],\n",
+      "[0.45898130536079407,],\n",
+      "[-0.21157969534397125,],\n",
+      "[1.0354582071304321,],\n",
+      "[0.8377259969711304,],\n",
+      "[-0.9411715269088745,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19389492273330688,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.13713476061820984,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09475503861904144,],\n",
+      "[0.45898130536079407,],\n",
+      "[-0.21157969534397125,],\n",
+      "[1.0354582071304321,],\n",
+      "[0.8377259969711304,],\n",
+      "[-0.9411715269088745,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.2608106771949679e-06,],\n",
+      "[-1.2152485169281135e-06,],\n",
+      "[-2.1072144591016695e-05,],\n",
+      "[8.882632755558006e-06,],\n",
+      "[-0.43410560488700867,],\n",
+      "[-0.006179526913911104,],\n",
+      "[-0.6311962604522705,],\n",
+      "[-0.24539847671985626,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19389492273330688,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.7406349778175354,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09518914669752121,],\n",
+      "[0.4589874744415283,],\n",
+      "[-0.21094849705696106,],\n",
+      "[1.0357036590576172,],\n",
+      "[0.8377743363380432,],\n",
+      "[-0.9411840438842773,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19315429031848907,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.0008557140245102346,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09518914669752121,],\n",
+      "[0.4589874744415283,],\n",
+      "[-0.21094849705696106,],\n",
+      "[1.0357036590576172,],\n",
+      "[0.8377743363380432,],\n",
+      "[-0.9411840438842773,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[7.280428349076828e-08,],\n",
+      "[-7.249803957165568e-08,],\n",
+      "[-1.324778281741601e-06,],\n",
+      "[5.750395644099626e-07,],\n",
+      "[-0.44676345586776733,],\n",
+      "[-0.006576988380402327,],\n",
+      "[-0.7212259769439697,],\n",
+      "[-0.24466083943843842,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19315429031848907,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.05850522965192795,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09563591331243515,],\n",
+      "[0.458994060754776,],\n",
+      "[-0.21022726595401764,],\n",
+      "[1.0359482765197754,],\n",
+      "[0.837822675704956,],\n",
+      "[-0.9411965608596802,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19309578835964203,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11809159070253372,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09563591331243515,],\n",
+      "[0.458994060754776,],\n",
+      "[-0.21022726595401764,],\n",
+      "[1.0359482765197754,],\n",
+      "[0.837822675704956,],\n",
+      "[-0.9411965608596802,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.251980835259019e-07,],\n",
+      "[6.4199883809124e-07,],\n",
+      "[1.2378923202049918e-05,],\n",
+      "[-5.536154276342131e-06,],\n",
+      "[-0.45885488390922546,],\n",
+      "[-0.0070509398356080055,],\n",
+      "[-0.8187482953071594,],\n",
+      "[-0.24460382759571075,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19309578835964203,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.6872891783714294,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.0960947647690773,],\n",
+      "[0.4590011239051819,],\n",
+      "[-0.20940852165222168,],\n",
+      "[1.0361928939819336,],\n",
+      "[0.8378710150718689,],\n",
+      "[-0.941209077835083,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19378307461738586,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.2703823745250702,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.0960947647690773,],\n",
+      "[0.4590011239051819,],\n",
+      "[-0.20940852165222168,],\n",
+      "[1.0361928939819336,],\n",
+      "[0.8378710150718689,],\n",
+      "[-0.941209077835083,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-6.91533898589114e-07,],\n",
+      "[7.307918963306292e-07,],\n",
+      "[1.4889592421241105e-05,],\n",
+      "[-6.865194336569402e-06,],\n",
+      "[-0.4635416865348816,],\n",
+      "[-0.007331117056310177,],\n",
+      "[-0.8609383702278137,],\n",
+      "[-0.2452911138534546,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19378307461738586,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[1.0399656295776367,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09655831009149551,],\n",
+      "[0.45900845527648926,],\n",
+      "[-0.20854757726192474,],\n",
+      "[1.036438226699829,],\n",
+      "[0.8379193544387817,],\n",
+      "[-0.9412215948104858,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19482304155826569,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.15285517275333405,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09655831009149551,],\n",
+      "[0.45900845527648926,],\n",
+      "[-0.20854757726192474,],\n",
+      "[1.036438226699829,],\n",
+      "[0.8379193544387817,],\n",
+      "[-0.9412215948104858,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-3.8008357705621165e-07,],\n",
+      "[4.1295672303931497e-07,],\n",
+      "[8.896585313777905e-06,],\n",
+      "[-4.230267677485244e-06,],\n",
+      "[-0.45896172523498535,],\n",
+      "[-0.007265132386237383,],\n",
+      "[-0.8248808979988098,],\n",
+      "[-0.24633106589317322,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19482304155826569,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.7819349765777588,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09701727330684662,],\n",
+      "[0.45901572704315186,],\n",
+      "[-0.20772269368171692,],\n",
+      "[1.036684513092041,],\n",
+      "[0.8379676938056946,],\n",
+      "[-0.9412341117858887,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.19560497999191284,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.001087940763682127,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09701727330684662,],\n",
+      "[0.45901572704315186,],\n",
+      "[-0.20772269368171692,],\n",
+      "[1.036684513092041,],\n",
+      "[0.8379676938056946,],\n",
+      "[-0.9412341117858887,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[-2.343993088516072e-08,],\n",
+      "[2.6201785630064478e-08,],\n",
+      "[5.965808327346167e-07,],\n",
+      "[-2.924781483670813e-07,],\n",
+      "[-0.4487413763999939,],\n",
+      "[-0.0069082132540643215,],\n",
+      "[-0.7385655641555786,],\n",
+      "[-0.2471114993095398,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.19560497999191284,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[0.06596784293651581,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09746601432561874,],\n",
+      "[0.4590226411819458,],\n",
+      "[-0.20698413252830505,],\n",
+      "[1.0369316339492798,],\n",
+      "[0.8380160331726074,],\n",
+      "[-0.9412466287612915,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.195670947432518,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "loss_antes tensor([0.11488907784223557,], device=\"cpu\", requires_grad=True)\n",
+      "f1 antes tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09746601432561874,],\n",
+      "[0.4590226411819458,],\n",
+      "[-0.20698413252830505,],\n",
+      "[1.0369316339492798,],\n",
+      "[0.8380160331726074,],\n",
+      "[-0.9412466287612915,]], device=\"cpu\", requires_grad=True)\n",
+      "f1 grad_antes tensor([[1.7607962377041986e-07,],\n",
+      "[-2.029000540915149e-07,],\n",
+      "[-4.875874310528161e-06,],\n",
+      "[2.463153350618086e-06,],\n",
+      "[-0.439397394657135,],\n",
+      "[-0.006496804300695658,],\n",
+      "[-0.6538427472114563,],\n",
+      "[-0.24717746675014496,],\n",
+      "[-0.04836709797382355,],\n",
+      "[0.012506773695349693,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 antes tensor([[-0.195670947432518,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 grad_antes tensor([[-0.6779057383537292,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "f1 depois tensor([[0.6356977224349976,],\n",
+      "[-0.12492786347866058,],\n",
+      "[0.5571243166923523,],\n",
+      "[-1.001272201538086,],\n",
+      "[0.09790541231632233,],\n",
+      "[0.4590291380882263,],\n",
+      "[-0.2063302844762802,],\n",
+      "[1.0371787548065186,],\n",
+      "[0.8380643725395203,],\n",
+      "[-0.9412591457366943,]], device=\"cpu\", requires_grad=True)\n",
+      "f2 depois tensor([[-0.1949930489063263,]], device=\"cpu\", requires_grad=True)\n",
+      "\n",
+      "\n",
+      "\n",
+      "Epoch [10/10], Loss: 0.1149\n"
      ]
     }
    ],
@@ -87,21 +20034,36 @@
     "for x in x_values:\n",
     "    y_true.append(math.pow(math.sin(x), 2))\n",
     "\n",
+    "batch_size = 100\n",
+    "\n",
     "\n",
     "for epoch in range(epochs):\n",
-    "    for x, target in zip(x_values, y_true):\n",
-    "        x = norch.Tensor([[x]]).T\n",
-    "        target = norch.Tensor([[target]]).T\n",
+    "    for i, (x, target) in enumerate(zip(x_values, y_true)):\n",
+    "        x = norch.Tensor([[x] for _ in range(batch_size)]).T\n",
+    "        target = norch.Tensor([[target] for _ in range(batch_size)]).T\n",
     "\n",
     "        x = x.to(device)\n",
     "        target = target.to(device)\n",
     "\n",
     "        outputs = model(x)\n",
+    "\n",
     "        loss = criterion(outputs, target)\n",
     "        \n",
     "        optimizer.zero_grad()\n",
     "        loss.backward()\n",
+    "        print('loss_antes', loss)\n",
+    "\n",
+    "        print('f1 antes', model.fc1.bias)\n",
+    "        print('f1 grad_antes', model.fc1.bias.grad)\n",
+    "        print('f2 antes', model.fc2.bias)\n",
+    "        print('f2 grad_antes', model.fc2.bias.grad)\n",
+    "\n",
     "        optimizer.step()\n",
+    "        print('\\n')\n",
+    "\n",
+    "        print('f1 depois', model.fc1.bias)\n",
+    "        print('f2 depois', model.fc2.bias)\n",
+    "        print('\\n\\n')\n",
     "\n",
     "    print(f'Epoch [{epoch + 1}/{epochs}], Loss: {loss[0]:.4f}')\n",
     "    loss_list.append(loss[0])"
@@ -109,28 +20071,66 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 9,
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "MyModel(\n",
-       "   (fc1): Linear(input_dim=1, output_dim=10, bias=True)\n",
-       "   (sigmoid): Sigmoid()\n",
-       "   (fc2): Linear(input_dim=10, output_dim=1, bias=True)\n",
-       ")"
+       "[1.7061519622802734]"
       ]
      },
-     "execution_count": 3,
+     "execution_count": 9,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
+   "source": [
+    "loss_list"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "ename": "NameError",
+     "evalue": "name 'model' is not defined",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
+      "Cell \u001b[0;32mIn[2], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[43mmodel\u001b[49m\n",
+      "\u001b[0;31mNameError\u001b[0m: name 'model' is not defined"
+     ]
+    }
+   ],
    "source": [
     "model"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "tensor([[-1.4443700313568115,]], device=\"cpu\", requires_grad=True)\n",
+      "0.1516466453264173\n"
+     ]
+    }
+   ],
+   "source": [
+    "x = 0.4\n",
+    "input = norch.Tensor([[x]]).T\n",
+    "print(model(input))\n",
+    "print(math.pow(math.sin(x), 2))"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -140,12 +20140,12 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [
     {
      "data": {
-      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA04AAAHWCAYAAABACtmGAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuNSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/xnp5ZAAAACXBIWXMAAA9hAAAPYQGoP6dpAABPhElEQVR4nO3deVxVdeL/8fcFBNzAHVFJXHJPRE0js7Ioc8s119J0mr6Vlea0aKWWTZk17ZqmU2mGqGNu5UqUtjm54miZaWqSCmomICYqnN8f58dVYkfgc+/l9Xw8zoPD4Rx43zNM3jfncz7HYVmWJQAAAABArrxMBwAAAAAAV0dxAgAAAIB8UJwAAAAAIB8UJwAAAADIB8UJAAAAAPJBcQIAAACAfFCcAAAAACAfFCcAAAAAyAfFCQAAAADyQXECgDLi3nvvVWhoaJGOfe655+RwOIo3kIcJDQ1Vz549Tccw6tChQ3I4HPrXv/5lOgoAFDuKEwAY5nA4CrRs2LDBdFQj7r33XlWqVMl0DJeQWUxyW15++WXTEQHAY/mYDgAAZd38+fOzfP7RRx8pJiYm2/bmzZtf0c+ZM2eOMjIyinTss88+q/Hjx1/Rz0fxGTJkiLp3755te3h4uIE0AFA2UJwAwLC77747y+f//e9/FRMTk237X509e1YVKlQo8M8pV65ckfJJko+Pj3x8+CfDVbRt2zbf3w8AQPFiqB4AuIGbb75ZrVq10rZt23TjjTeqQoUKevrppyVJK1asUI8ePVSnTh35+fmpUaNGeuGFF5Senp7le/z1HqfL70eZPXu2GjVqJD8/P1177bXasmVLlmNzusfJ4XDo4Ycf1vLly9WqVSv5+fmpZcuWWrt2bbb8GzZsUPv27eXv769GjRrpvffeK/b7pv7zn/+oXbt2Kl++vGrUqKG7775bR44cybJPQkKCRo4cqXr16snPz0/BwcHq3bu3Dh065Nxn69at6tq1q2rUqKHy5curQYMGGjVqVIFzrF+/Xm3atJG/v79atGihpUuXOr924MABORwOvfHGG9mO++677+RwOBQdHV34F5+DzHuu8spzea677rpL1apVU4UKFXTddddp1apV2fY7d+6cnnvuOTVp0kT+/v4KDg5Wv3799Msvv2TbN7/fKQBwN/z5EADcxO+//65u3bpp8ODBuvvuuxUUFCRJmjt3ripVqqRx48apUqVK+uKLLzRp0iQlJyfr1Vdfzff7LliwQCkpKfq///s/ORwOvfLKK+rXr58OHDiQ71Wqb775RkuXLtVDDz2kypUr6+2331b//v11+PBhVa9eXZK0Y8cO3XHHHQoODtbzzz+v9PR0TZkyRTVr1rzyk/L/zZ07VyNHjtS1116rqVOnKjExUW+99Za+/fZb7dixQ1WqVJEk9e/fXz/88IMeeeQRhYaG6vjx44qJidHhw4edn99+++2qWbOmxo8frypVqujQoUM5lo2c7Nu3T4MGDdIDDzygESNG6MMPP9Rdd92ltWvX6rbbblPDhg3VqVMnRUVF6bHHHstybFRUlCpXrqzevXvn+3POnj2rkydPZttepUqVLFcG88sjSYmJibr++ut19uxZPfroo6pevbrmzZunO++8U0uWLFHfvn0lSenp6erZs6diY2M1ePBgjRkzRikpKYqJidHu3bvVqFEj58+9kt8pAHBZFgDApYwePdr663+eb7rpJkuSNWvWrGz7nz17Ntu2//u//7MqVKhgnTt3zrltxIgRVv369Z2fHzx40JJkVa9e3Tp16pRz+4oVKyxJ1qeffurcNnny5GyZJFm+vr7W/v37ndt27txpSbLeeecd57ZevXpZFSpUsI4cOeLctm/fPsvHxyfb98zJiBEjrIoVK+b69fPnz1u1atWyWrVqZf3555/O7Z999pklyZo0aZJlWZb1xx9/WJKsV199NdfvtWzZMkuStWXLlnxz/VX9+vUtSdYnn3zi3JaUlGQFBwdb4eHhzm3vvfeeJcnas2dPltdQo0YNa8SIEXn+jMz/zXJbNm3aVOg8Y8eOtSRZX3/9tXNbSkqK1aBBAys0NNRKT0+3LMuyPvjgA0uS9frrr2fLlZGRkSVfQX6nAMDdMFQPANyEn5+fRo4cmW17+fLlnespKSk6efKkOnfurLNnz+qnn37K9/sOGjRIVatWdX7euXNnSfbwrfxERkZmudLQunVrBQQEOI9NT0/X559/rj59+qhOnTrO/Ro3bqxu3brl+/0LYuvWrTp+/Lgeeugh+fv7O7f36NFDzZo1cw45K1++vHx9fbVhwwb98ccfOX6vzCtTn332mS5cuFDoLHXq1HFeoZGkgIAADR8+XDt27FBCQoIkaeDAgfL391dUVJRzv3Xr1unkyZMFvm/p/vvvV0xMTLalRYsWhc6zevVqdejQQTfccINzv0qVKun+++/XoUOH9OOPP0qSPvnkE9WoUUOPPPJItjx/HXJ5Jb9TAOCqKE4A4Cbq1q0rX1/fbNt/+OEH9e3bV4GBgQoICFDNmjWdb8CTkpLy/b5XXXVVls8z3/DmVi7yOjbz+Mxjjx8/rj///FONGzfOtl9O24ri119/lSQ1bdo029eaNWvm/Lqfn5+mTZumNWvWKCgoSDfeeKNeeeUVZ4GQpJtuukn9+/fX888/rxo1aqh379768MMPlZaWVqAsjRs3zlYimjRpIknO+6iqVKmiXr16acGCBc59oqKiVLduXd1yyy0F+jlXX321IiMjsy0BAQGFzvPrr7/meO4yZ3HMPH+//PKLmjZtWqBJQq7kdwoAXBXFCQDcxOVXljKdPn1aN910k3bu3KkpU6bo008/VUxMjKZNmyZJBZp+3NvbO8ftlmWV6LEmjB07Vj///LOmTp0qf39/TZw4Uc2bN9eOHTsk2VdOlixZok2bNunhhx/WkSNHNGrUKLVr105nzpwpthzDhw/XgQMH9N133yklJUUrV67UkCFD5OXlGf8su9vvBQAUhGf8FxoAyqgNGzbo999/19y5czVmzBj17NlTkZGRWYZJmVSrVi35+/tr//792b6W07aiqF+/viRp79692b62d+9e59czNWrUSP/4xz+0fv167d69W+fPn9drr72WZZ/rrrtOL774orZu3aqoqCj98MMPWrhwYb5Z9u/fn60c/Pzzz5KUZUbDO+64QzVr1lRUVJSWLVums2fP6p577inQ6y2MguSpX79+jucuc5hn5vlr1KiR9u7dW6QhjADgCShOAODGMv+yf/mb4/Pnz+vdd981FSkLb29vRUZGavny5Tp69Khz+/79+7VmzZpi+Rnt27dXrVq1NGvWrCxD6tasWaM9e/aoR48ekuyZ6M6dO5fl2EaNGqly5crO4/74449sRaNNmzaSVKDhekePHtWyZcucnycnJ+ujjz5SmzZtVLt2bed2Hx8fDRkyRIsXL9bcuXN1zTXXqHXr1oV74QVQkDzdu3fX5s2btWnTJud+qampmj17tkJDQ533TfXv318nT57U9OnTs/0criQBKAuYjhwA3Nj111+vqlWrasSIEXr00UflcDg0f/58l3oj+9xzz2n9+vXq1KmTHnzwQaWnp2v69Olq1aqV4uLiCvQ9Lly4oH/+85/ZtlerVk0PPfSQpk2bppEjR+qmm27SkCFDnNORh4aGOqf9/vnnn3Xrrbdq4MCBatGihXx8fLRs2TIlJiZq8ODBkqR58+bp3XffVd++fdWoUSOlpKRozpw5CggIUPfu3fPN2aRJE/3tb3/Tli1bFBQUpA8++ECJiYn68MMPs+07fPhwvf322/ryyy+dQysLavv27fr444+zbW/UqJEiIiIKlWf8+PGKjo5Wt27d9Oijj6patWqaN2+eDh48qE8++cQ5fHD48OH66KOPNG7cOG3evFmdO3dWamqqPv/8cz300EMFmkYdANwZxQkA3Fj16tX12Wef6R//+IeeffZZVa1aVXfffbduvfVWde3a1XQ8SVK7du20Zs0aPf7445o4caJCQkI0ZcoU7dmzp0Cz/kn2VbSJEydm296oUSM99NBDuvfee1WhQgW9/PLLeuqpp1SxYkX17dtX06ZNc86UFxISoiFDhig2Nlbz58+Xj4+PmjVrpsWLF6t///6S7MkhNm/erIULFyoxMVGBgYHq0KGDoqKi1KBBg3xzXn311XrnnXf0xBNPaO/evWrQoIEWLVqU4/8W7dq1U8uWLbVnzx4NGzasQOchU3R0dI4Pyh0xYkSW4lSQPEFBQfruu+/01FNP6Z133tG5c+fUunVrffrpp86rdZJ99XD16tV68cUXtWDBAn3yySeqXr26brjhBl1zzTWFyg8A7shhudKfJQEAZUafPn30ww8/aN++faajGBMeHq5q1aopNja22L93aGioWrVqpc8++6zYvzcAlEXc4wQAKHF//vlnls/37dun1atX6+abbzYTyAVs3bpVcXFxGj58uOkoAIACYKgeAKDENWzYUPfee68aNmyoX3/9VTNnzpSvr6+efPJJ09FK3e7du7Vt2za99tprCg4O1qBBg0xHAgAUAMUJAFDi7rjjDkVHRyshIUF+fn6KiIjQSy+9pKuvvtp0tFK3ZMkSTZkyRU2bNlV0dLT8/f1NRwIAFAD3OAEAAABAPrjHCQAAAADyQXECAAAAgHyUuXucMjIydPToUVWuXFkOh8N0HAAAAACGWJallJQU1alTx/nA79yUueJ09OhRhYSEmI4BAAAAwEXEx8erXr16ee5T5opT5cqVJdknJyAgwHAaAAAAAKYkJycrJCTE2RHyUuaKU+bwvICAAIoTAAAAgALdwsPkEAAAAACQD4oTAAAAAOSD4gQAAAAA+aA4AQAAAEA+KE4AAAAAkA+KEwAAAADkg+IEAAAAAPmgOAEAAABAPihOAAAAAJAPH9MByrL0dOnrr6Vjx6TgYKlzZ8nb23QqAAAAAH9FcTJk6VJpzBjpt98ubatXT3rrLalfP3O5AAAAAGTHUD0Dli6VBgzIWpok6cgRe/vSpWZyAQAAAMgZxamUpafbV5osK/vXMreNHWvvBwAAAMA1UJxK2ddfZ7/SdDnLkuLj7f0AAAAAuAaKUyk7dqx49wMAAABQ8ihOpSw4uHj3AwAAAFDyKE6lrHNne/Y8hyPnrzscUkiIvR8AAAAA10BxKmXe3vaU41L28pT5+Ztv8jwnAAAAwJVQnAzo109askSqWzfr9ho17O08xwkAAABwLRQnQ/r1kw4dkr78UrrtNnvb7bdTmgAAAABX5GM6QFnm7S3dfLPk5yfFxEjLl0upqVLFiqaTAQAAALgcV5xcwHXXSQ0b2qVpxQrTaQAAAAD8FcXJBTgc0t132+sff2w2CwAAAIDsKE4uYtgw++P69VJiotksAAAAALKiOLmIJk2kDh2k9HRp0SLTaQAAAABcjuLkQhiuBwAAALgmipMLGTTInmlvyxZp717TaQAAAABkoji5kFq1pK5d7fWoKLNZAAAAAFxCcXIxlw/XsyyzWQAAAADYKE4upndvqVIl6eBBadMm02kAAAAASBQnl1OhgtSvn73OJBEAAACAa6A4uaDM4XqLFknnz5vNAgAAAIDi5JJuuUWqXVs6dUpau9Z0GgAAAABGi9NXX32lXr16qU6dOnI4HFq+fHm+x6SlpemZZ55R/fr15efnp9DQUH3wwQclH7YUeXtLQ4fa6wzXAwAAAMwzWpxSU1MVFhamGTNmFPiYgQMHKjY2Vu+//7727t2r6OhoNW3atARTmpE5XG/lSikpyWwWAAAAoKzzMfnDu3Xrpm7duhV4/7Vr12rjxo06cOCAqlWrJkkKDQ0toXRmtWkjtWgh/fij9Mkn0qhRphMBAAAAZZdb3eO0cuVKtW/fXq+88orq1q2rJk2a6PHHH9eff/6Z6zFpaWlKTk7OsrgDhyPrM50AAAAAmONWxenAgQP65ptvtHv3bi1btkxvvvmmlixZooceeijXY6ZOnarAwEDnEhISUoqJr0zmfU4bNkjx8UajAAAAAGWaWxWnjIwMORwORUVFqUOHDurevbtef/11zZs3L9erThMmTFBSUpJziXejBlK/vnTjjZJlSdHRptMAAAAAZZdbFafg4GDVrVtXgYGBzm3NmzeXZVn67bffcjzGz89PAQEBWRZ3wnA9AAAAwDy3Kk6dOnXS0aNHdebMGee2n3/+WV5eXqpXr57BZCVnwADJ11fatUv63/9MpwEAAADKJqPF6cyZM4qLi1NcXJwk6eDBg4qLi9Phw4cl2cPshg8f7tx/6NChql69ukaOHKkff/xRX331lZ544gmNGjVK5cuXN/ESSlzVqlLPnvY6V50AAAAAM4wWp61btyo8PFzh4eGSpHHjxik8PFyTJk2SJB07dsxZoiSpUqVKiomJ0enTp9W+fXsNGzZMvXr10ttvv20kf2nJHK63YIGUnm42CwAAAFAWOSzLskyHKE3JyckKDAxUUlKS29zvlJYm1a4tnT4txcZKt9xiOhEAAADg/grTDdzqHqeyys9PGjjQXme4HgAAAFD6KE5uInO43pIlUh7P+wUAAABQAihObqJTJ/u5Tikp0qefmk4DAAAAlC0UJzfh5SUNG2avM1wPAAAAKF0UJzeSWZzWrJFOnjSbBQAAAChLKE5upEULqW1b6eJFafFi02kAAACAsoPi5GYyJ4lguB4AAABQeihObmbwYPt+p02bpF9+MZ0GAAAAKBsoTm4mOFiKjLTXo6LMZgEAAADKCoqTG7p8uJ5lmc0CAAAAlAUUJzfUt69UoYK0b5+0ZYvpNAAAAIDnozi5oUqVpD597HUmiQAAAABKHsXJTWUO11u4ULpwwWwWAAAAwNNRnNzUbbdJNWtKJ05IMTGm0wAAAACejeLkpnx8pCFD7HWG6wEAAAAli+LkxjKH6y1fLqWkGI0CAAAAeDSKkxtr315q0kT6809p2TLTaQAAAADPRXFyYw5H1mc6AQAAACgZFCc3N2yY/TE2Vjp61GwWAAAAwFNRnNxcw4bS9ddLGRn21OQAAAAAih/FyQMwXA8AAAAoWRQnDzBwoD09+Y4d0g8/mE4DAAAAeB6KkweoXl3q3t1ej4oymwUAAADwRBQnD5E5XC8qyr7fCQAAAEDxoTh5iJ49pYAA6fBh6ZtvTKcBAAAAPAvFyUOULy8NGGCvM0kEAAAAULwoTh4kc7je4sXSuXNmswAAAACehOLkQW66SapXT0pKklavNp0GAAAA8BwUJw/i5SUNHWqvM1wPAAAAKD4UJw+TOVxv1Srp1CmzWQAAAABPQXHyMNdcI7VuLZ0/Ly1ZYjoNAAAA4BkoTh4o86oTw/UAAACA4kFx8kBDhkgOh/T119KhQ6bTAAAAAO6P4uSB6tWTunSx1xcsMJsFAAAA8ARGi9NXX32lXr16qU6dOnI4HFq+fHmBj/3222/l4+OjNm3alFg+d5Y5XG/+fMmyzGYBAAAA3J3R4pSamqqwsDDNmDGjUMedPn1aw4cP16233lpCydxfv36Sv7/000/Sjh2m0wAAAADuzcfkD+/WrZu6detW6OMeeOABDR06VN7e3vlepUpLS1NaWprz8+Tk5EL/PHcUGCjdeae0eLE9SUTbtqYTAQAAAO7L7e5x+vDDD3XgwAFNnjy5QPtPnTpVgYGBziUkJKSEE7qOzOF60dHSxYtmswAAAADuzK2K0759+zR+/Hh9/PHH8vEp2MWyCRMmKCkpybnEx8eXcErX0bWrVL26lJAgffGF6TQAAACA+3Kb4pSenq6hQ4fq+eefV5MmTQp8nJ+fnwICArIsZYWvrzRokL3OM50AAACAonOb4pSSkqKtW7fq4Ycflo+Pj3x8fDRlyhTt3LlTPj4++oJLKjnKHK63dKmUmmo2CwAAAOCujE4OURgBAQHatWtXlm3vvvuuvvjiCy1ZskQNGjQwlMy1XXed1LChdOCAtGKFNHSo6UQAAACA+zFanM6cOaP9+/c7Pz948KDi4uJUrVo1XXXVVZowYYKOHDmijz76SF5eXmrVqlWW42vVqiV/f/9s23GJw2FfdZoyxR6uR3ECAAAACs/oUL2tW7cqPDxc4eHhkqRx48YpPDxckyZNkiQdO3ZMhw8fNhnRIwwbZn9cv15KTDSbBQAAAHBHDsuyLNMhSlNycrICAwOVlJRUpiaK6NhR2rxZeust6dFHTacBAAAAzCtMN3CbySFwZTIniWB2PQAAAKDwKE5lxKBBkre3tGWLtHev6TQAAACAe6E4lRG1atkPxJWkqCizWQAAAAB3Q3EqQy4frle27mwDAAAArgzFqQzp3VuqVEk6eFDatMl0GgAAAMB9UJzKkAoVpH797HUmiQAAAAAKjuJUxmQO11u0SDp/3mwWAAAAwF1QnMqYW26RateWTp2S1q41nQYAAABwDxSnMsbbWxo61F5nuB4AAABQMBSnMihzuN7KlVJSktksAAAAgDugOJVBbdpILVpIaWnSJ5+YTgMAAAC4PopTGeRwZH2mEwAAAIC8UZzKqMz7nDZskOLjjUYBAAAAXB7FqYyqX1+68UbJsqToaNNpAAAAANdGcSrDGK4HAAAAFAzFqQwbMEDy9ZV27ZL+9z/TaQAAAADXRXEqw6pWlXr2tNe56gQAAADkjuJUxmUO11uwQEpPN5sFAAAAcFUUpzKue3epShXpyBFp40bTaQAAAADXRHEq4/z8pIED7XWG6wEAAAA5ozjBOVxvyRLpzz/NZgEAAABcEcUJ6tTJfq5TSor06aem0wAAAACuh+IEeXlJw4bZ6wzXAwAAALKjOEHSpeK0Zo108qTZLAAAAICroThBktSihdS2rXTxorR4sek0AAAAgGuhOMEpc5IIhusBAAAAWVGc4DR4sH2/06ZN0i+/mE4DAAAAuA6KE5yCg6XISHs9KspsFgAAAMCVUJyQxT332B8//liyLLNZAAAAAFdBcUIWffpIFSpI+/ZJW7aYTgMAAAC4BooTsqhUSerb115nkggAAADARnFCNpmz6y1cKF24YDYLAAAA4AooTsgmMlKqVUs6cUKKiTGdBgAAADDPaHH66quv1KtXL9WpU0cOh0PLly/Pc/+lS5fqtttuU82aNRUQEKCIiAitW7eudMKWIT4+0pAh9jrD9QAAAADDxSk1NVVhYWGaMWNGgfb/6quvdNttt2n16tXatm2bunTpol69emnHjh0lnLTsyRyut3y5lJJiNAoAAABgnMOyXGPSaYfDoWXLlqlPnz6FOq5ly5YaNGiQJk2aVKD9k5OTFRgYqKSkJAUEBBQhadlgWVLz5tLevdK8edLw4aYTAQAAAMWrMN3Are9xysjIUEpKiqpVq5brPmlpaUpOTs6yIH8Ox6WrTgzXAwAAQFnn1sXpX//6l86cOaOBAwfmus/UqVMVGBjoXEJCQkoxoXsbOtT+GBsrHT1qNgsAAABgktsWpwULFuj555/X4sWLVatWrVz3mzBhgpKSkpxLfHx8KaZ0bw0bSp06SRkZ9tTkAAAAQFnllsVp4cKFuu+++7R48WJFRkbmua+fn58CAgKyLCg4husBAAAAblicoqOjNXLkSEVHR6tHjx6m43i8u+6SypWTduyQfvjBdBoAAADADKPF6cyZM4qLi1NcXJwk6eDBg4qLi9Phw4cl2cPshl82nduCBQs0fPhwvfbaa+rYsaMSEhKUkJCgpKQkE/HLhOrVpe7d7fWoKLNZAAAAAFOMFqetW7cqPDxc4eHhkqRx48YpPDzcObX4sWPHnCVKkmbPnq2LFy9q9OjRCg4Odi5jxowxkr+syByuFxVl3+8EAAAAlDUu8xyn0sJznArv3DkpKEhKTpY2bpRuvNF0IgAAAODKlZnnOKF0+Pvb9zpJTBIBAACAsonihALJHK63eLF9BQoAAAAoSyhOKJAbb5Tq1ZOSkqTVq02nAQAAAEoXxQkF4uUlDRtmrzNcDwAAAGUNxQkFljlcb9Uq6dQps1kAAACA0kRxQoG1aiWFhUnnz0tLlphOAwAAAJQeihMKJfOqE8P1AAAAUJZQnFAoQ4ZIDof09dfSoUOm0wAAAAClg+KEQqlbV7rlFnt9wQKzWQAAAIDSQnFCoWUO15s/X7Iss1kAAACA0kBxQqH16yf5+0s//STt2GE6DQAAAFDyKE4otIAAqXdve51JIgAAAFAWUJxQJJnD9aKjpYsXzWYBAAAAShrFCUXStatUvbqUkCB98YXpNAAAAEDJojihSMqVkwYPttcZrgcAAABPR3FCkWUO11u6VEpNNZsFAAAAKEkUJxRZx45So0Z2aVqxwnQaAAAAoORQnFBkDselq04M1wMAAIAnozjhigwbZn9cv15KTDSbBQAAACgpFCdckauvtofspadLixaZTgMAAACUDIoTrhjD9QAAAODpKE64YoMGSd7e0pYt0t69ptMAAAAAxY/ihCtWs6Z0xx32elSU2SwAAABASaA4oVhcPlzPssxmAQAAAIobxQnF4s47pUqVpIMHpU2bTKcBAAAAihfFCcWiQgWpf397nUkiAAAA4GkoTig2mcP1Fi2Szp83mwUAAAAoThQnFJsuXaTgYOnUKWntWtNpAAAAgOJDcUKx8faWhg611xmuBwAAAE9CcUKxyhyut3KllJRkNgsAAABQXChOKFZhYVLLllJamvTJJ6bTAAAAAMWD4oRi5XBkfaYTAAAA4AkoTih2mfc5bdggxccbjQIAAAAUC6PF6auvvlKvXr1Up04dORwOLV++PN9jNmzYoLZt28rPz0+NGzfW3LlzSzwnCueqq6SbbpIsS4qONp0GAAAAuHJGi1NqaqrCwsI0Y8aMAu1/8OBB9ejRQ126dFFcXJzGjh2r++67T+vWrSvhpCgshusBAADAkzgsy7JMh5Akh8OhZcuWqU+fPrnu89RTT2nVqlXavXu3c9vgwYN1+vRprS3gg4OSk5MVGBiopKQkBQQEXGls5OL0aSkoyH4Q7s6dUuvWphMBAAAAWRWmG7jVPU6bNm1SZGRklm1du3bVpk2bcj0mLS1NycnJWRaUvCpVpF697HWuOgEAAMDduVVxSkhIUFBQUJZtQUFBSk5O1p9//pnjMVOnTlVgYKBzCQkJKY2o0KXhegsWSOnpZrMAAAAAV8KtilNRTJgwQUlJSc4lnmneSk23blLVqtKRI9LGjabTAAAAAEXnVsWpdu3aSkxMzLItMTFRAQEBKl++fI7H+Pn5KSAgIMuC0uHnJw0caK8zXA8AAADuzK2KU0REhGJjY7Nsi4mJUUREhKFEyE/mcL0lS6RcRlMCAAAALs9ocTpz5ozi4uIUFxcnyZ5uPC4uTocPH5ZkD7MbPny4c/8HHnhABw4c0JNPPqmffvpJ7777rhYvXqzHHnvMRHwUwPXXS6GhUkqK9OmnptMAAAAARWO0OG3dulXh4eEKDw+XJI0bN07h4eGaNGmSJOnYsWPOEiVJDRo00KpVqxQTE6OwsDC99tpr+ve//62uXbsayY/8eXlJw4bZ6wzXAwAAgLtymec4lRae41T69uyRWrSQfHykY8ekGjVMJwIAAAA8+DlOcE/Nm0vt2kkXL0qLF5tOAwAAABRekYpTfHy8fvvtN+fnmzdv1tixYzV79uxiCwbPkjlJBMP1AAAA4I6KVJyGDh2qL7/8UpL9UNrbbrtNmzdv1jPPPKMpU6YUa0B4hsGD7fudNm2SfvnFdBoAAACgcIpUnHbv3q0OHTpIkhYvXqxWrVrpu+++U1RUlObOnVuc+eAhateWbrvNXo+KMpsFAAAAKKwiFacLFy7Iz89PkvT555/rzjvvlCQ1a9ZMx44dK7508CiZw/Xmz5fK1pQkAAAAcHdFKk4tW7bUrFmz9PXXXysmJkZ33HGHJOno0aOqXr16sQaE5+jTR6pQQdq/X9q82XQaAAAAoOCKVJymTZum9957TzfffLOGDBmisLAwSdLKlSudQ/iAv6pUSerb115nkggAAAC4kyI/xyk9PV3JycmqWrWqc9uhQ4dUoUIF1apVq9gCFjee42TW2rVSt272s5yOHpXKlTOdCAAAAGVViT/H6c8//1RaWpqzNP3666968803tXfvXpcuTTAvMlKqVUs6eVJav950GgAAAKBgilScevfurY8++kiSdPr0aXXs2FGvvfaa+vTpo5kzZxZrQHgWHx9pyBB7neF6AAAAcBdFKk7bt29X586dJUlLlixRUFCQfv31V3300Ud6++23izUgPE/m7HrLl0vJyUajAAAAAAVSpOJ09uxZVa5cWZK0fv169evXT15eXrruuuv066+/FmtAeJ527aSmTaVz56Rly0ynAQAAAPJXpOLUuHFjLV++XPHx8Vq3bp1uv/12SdLx48eZcAH5cjguXXViuB4AAADcQZGK06RJk/T4448rNDRUHTp0UEREhCT76lN4eHixBoRnGjrU/hgba8+uBwAAALiyIhWnAQMG6PDhw9q6davWrVvn3H7rrbfqjTfeKLZw8FwNG0qdOkmWJUVHm04DAAAA5K1IxUmSateurfDwcB09elS//fabJKlDhw5q1qxZsYWDZ2O4HgAAANxFkYpTRkaGpkyZosDAQNWvX1/169dXlSpV9MILLygjI6O4M8JD3XWX/QDcuDhp927TaQAAAIDcFak4PfPMM5o+fbpefvll7dixQzt27NBLL72kd955RxMnTizujPBQ1atL3bvb61FRZrMAAAAAeXFYlmUV9qA6depo1qxZuvPOO7NsX7FihR566CEdOXKk2AIWt+TkZAUGBiopKYkZAF3AkiX2laeQEOnQIcmryINHAQAAgMIpTDco0tvUU6dO5XgvU7NmzXTq1KmifEuUUT17SgEBUny89PXXptMAAAAAOStScQoLC9P06dOzbZ8+fbpat259xaFQdvj721ecJCaJAAAAgOsq0lC9jRs3qkePHrrqqqucz3DatGmT4uPjtXr1anXu3LnYgxYXhuq5ng0bpC5dpMBAKSHBLlMAAABASSvxoXo33XSTfv75Z/Xt21enT5/W6dOn1a9fP/3www+aP39+kUKj7LrxRqlePSkpSVq1ynQaAAAAILsiXXHKzc6dO9W2bVulp6cX17csdlxxck3jx0vTpkl9+kjLlplOAwAAgLKgxK84AcUt82G4q1ZJzC8CAAAAV0Nxgkto1UoKC5MuXJD+8x/TaQAAAICsKE5wGZlXnZhdDwAAAK7GpzA79+vXL8+vnz59+kqyoIwbMkR68knpm2+kgwelBg1MJwIAAABshbriFBgYmOdSv359DR8+vKSywsPVrSvdcou9vmCB2SwAAADA5Yp1Vj13wKx6rm3uXGnkSKlZM+nHHyWHw3QiAAAAeCpm1YPb6tfPfgDuTz9J27ebTgMAAADYKE5wKQEBUu/e9jqTRAAAAMBVUJzgcjJn14uOli5eNJsFAAAAkFykOM2YMUOhoaHy9/dXx44dtXnz5jz3f/PNN9W0aVOVL19eISEheuyxx3Tu3LlSSouS1rWrVL26lJgoxcaaTgMAAAC4QHFatGiRxo0bp8mTJ2v79u0KCwtT165ddfz48Rz3X7BggcaPH6/Jkydrz549ev/997Vo0SI9/fTTpZwcJaVcOWnwYHud4XoAAABwBcaL0+uvv66///3vGjlypFq0aKFZs2apQoUK+uCDD3Lc/7vvvlOnTp00dOhQhYaG6vbbb9eQIUPyvUoF95I5XG/pUunMGbNZAAAAAKPF6fz589q2bZsiIyOd27y8vBQZGalNmzbleMz111+vbdu2OYvSgQMHtHr1anXv3j3H/dPS0pScnJxlgevr2FFq1Eg6e1ZascJ0GgAAAJR1RovTyZMnlZ6erqCgoCzbg4KClJCQkOMxQ4cO1ZQpU3TDDTeoXLlyatSokW6++eZch+pNnTo1y0N6Q0JCiv11oPg5HJeuOjFcDwAAAKYZH6pXWBs2bNBLL72kd999V9u3b9fSpUu1atUqvfDCCznuP2HCBCUlJTmX+Pj4Uk6Moho2zP64fr09UQQAAABgio/JH16jRg15e3sr8S/vihMTE1W7du0cj5k4caLuuece3XfffZKka665Rqmpqbr//vv1zDPPyMsraxf08/OTn59fybwAlKirr7aH7H3/vbRwoTRmjOlEAAAAKKuMXnHy9fVVu3btFHvZnNMZGRmKjY1VREREjsecPXs2Wzny9vaWJFmWVXJhYQTD9QAAAOAKjA/VGzdunObMmaN58+Zpz549evDBB5WamqqRI0dKkoYPH64JEyY49+/Vq5dmzpyphQsX6uDBg4qJidHEiRPVq1cvZ4GC5xg0SPL2lrZulX76yXQaAAAAlFVGh+pJ0qBBg3TixAlNmjRJCQkJatOmjdauXeucMOLw4cNZrjA9++yzcjgcevbZZ3XkyBHVrFlTvXr10osvvmjqJaAE1awp3XGHtGqVFBUl5XIrGwAAAFCiHFYZG9+WnJyswMBAJSUlKSAgwHQcFMDChdKQIVJoqHTggD3jHgAAAHClCtMNjA/VA/Jz551SpUrSoUPSd9+ZTgMAAICyiOIEl1ehgtS/v73OJBEAAAAwgeIEt5A5u96iRdL582azAAAAoOyhOMEtdOkiBQdLf/whrVljOg0AAADKGooT3IK3tzR0qL3OcD0AAACUNooT3EbmcL1PP5VOnzYaBQAAAGUMxQluIyxMatlSSkuTXnxRio6WNmyQ0tNNJwMAAICnM/4AXKCgHA6pTRvphx+kf/3r0vZ69aS33pL69TMWDQAAAB6OK05wG0uXSgsWZN9+5Ig0YID9dQAAAKAkUJzgFtLTpTFjJMvK/rXMbWPHMmwPAAAAJYPiBLfw9dfSb7/l/nXLkuLj7f0AAACA4kZxgls4dqx49wMAAAAKg+IEtxAcXLz7AQAAAIVBcYJb6NzZnj3P4ch9n5AQez8AAACguFGc4Ba8ve0px6Xcy9PIkfZ+AAAAQHGjOMFt9OsnLVki1a2bdXuFCvbHN9+Udu8u9VgAAAAoAyhOcCv9+kmHDklffmk/0+nLL6UTJ+whesnJUs+eUkKC6ZQAAADwND6mAwCF5e0t3Xxz1m3LlkkREdK+fdKdd0obNly6EgUAAABcKa44wSNUry6tWiVVqyZt2SLdc4+UkWE6FQAAADwFxQke4+qrpeXLJV9faelSafx404kAAADgKShO8CidO0sffGCvv/qqNHu22TwAAADwDBQneJxhw6Tnn7fXH3pIWr/ebB4AAAC4P4oTPNLEifZ9Tunp0oABTFMOAACAK0NxgkdyOKQ5c6Qbb5RSUqQePZimHAAAAEVHcYLH8vOzpylv0kQ6fNiepvzsWdOpAAAA4I4oTvBo1arZ05RXr25PU3733UxTDgAAgMKjOMHjNW58aZryZcukp54ynQgAAADuhuKEMuGGG6QPP7TX//Uv6b33zOYBAACAe6E4ocwYOlSaMsVeHz1aWrfObB4AAAC4D4oTypRnn5WGD7enKb/rLmnXLtOJAAAA4A4oTihTHA5p9mzpppvsacp79pSOHTOdCgAAAK6O4oQyx89PWro06zTlqammUwEAAMCVUZxQJl0+TfnWrdI999jD9wAAAICcUJxQZjVuLK1YwTTlAAAAyJ9LFKcZM2YoNDRU/v7+6tixozZv3pzn/qdPn9bo0aMVHBwsPz8/NWnSRKtXry6ltPAknTpJc+fa66+9Js2aZTQOAAAAXJTx4rRo0SKNGzdOkydP1vbt2xUWFqauXbvq+PHjOe5//vx53XbbbTp06JCWLFmivXv3as6cOapbt24pJ4enGDJEeuEFe/3hh5mmHAAAANk5LMuyTAbo2LGjrr32Wk2fPl2SlJGRoZCQED3yyCMaP358tv1nzZqlV199VT/99JPKlSuX7/dPS0tTWlqa8/Pk5GSFhIQoKSlJAQEBxfdC4NYsSxo5Upo3T6pcWfr2W+maa0ynAgAAQElKTk5WYGBggbqB0StO58+f17Zt2xQZGenc5uXlpcjISG3atCnHY1auXKmIiAiNHj1aQUFBatWqlV566SWl53Jn/9SpUxUYGOhcQkJCSuS1wL1lTlN+8832NOU9ejBNOQAAAC4xWpxOnjyp9PR0BQUFZdkeFBSkhISEHI85cOCAlixZovT0dK1evVoTJ07Ua6+9pn/+85857j9hwgQlJSU5l/j4+GJ/HfAMvr7SJ5/Y05THx0u9ejFNOQAAAGw+pgMUVkZGhmrVqqXZs2fL29tb7dq105EjR/Tqq69q8uTJ2fb38/OTn5+fgaRwR9WqSatXS9ddJ23bJg0bZpcpb2/TyQAAAGCS0StONWrUkLe3txITE7NsT0xMVO3atXM8Jjg4WE2aNJH3Ze9kmzdvroSEBJ0/f75E86JsaNRIWr7cflDuihXSk0+aTgQAAADTjBYnX19ftWvXTrGxsc5tGRkZio2NVURERI7HdOrUSfv371dGRoZz288//6zg4GD5+vqWeGaUDZdPU/7669LMmUbjAAAAwDDj05GPGzdOc+bM0bx587Rnzx49+OCDSk1N1ciRIyVJw4cP14QJE5z7P/jggzp16pTGjBmjn3/+WatWrdJLL72k0aNHm3oJ8FCDB0uZt8498oi0dq3ZPAAAADDH+D1OgwYN0okTJzRp0iQlJCSoTZs2Wrt2rXPCiMOHD8vL61K/CwkJ0bp16/TYY4+pdevWqlu3rsaMGaOnnnrK1EuAB3v6aWnfPnua8oEDpW++kVq3Np0KAAAApc34c5xKW2Hmagck6fx5qWtXacMGKSRE+v57KTjYdCoAAABcKbd5jhPgDnx9paVLpaZNmaYcAACgrKI4AQVQtaq0apVUo8alacpzeeYyAAAAPBDFCSigRo3s6ckzpyl/4gnTiQAAAFBaKE5AIVx//aVpyt94Q3r3XaNxAAAAUEooTkAhDR4svfiivf7II9KaNWbzAAAAoORRnIAimDBBuvdeKSPDnqZ8507TiQAAAFCSKE5AETgc0nvvSV26SGfOSD17SkePmk4FAACAkkJxAorI11f65BOpWTPpt9+YphwAAMCTUZyAK5A5TXnNmtL27dLQoUxTDgAA4IkoTsAVatjw0jTlK1cyTTkAAIAnojgBxSAiQpo3z15/4w1pxgyzeQAAAFC8KE5AMRk0SHrpJXv90Uel1avN5gEAAEDxoTgBxWj8eGnUKHua8kGDmKYcAADAU1CcgGLkcEgzZ0q33MI05QAAAJ6E4gQUM19facmSS9OU9+xplygAAAC4L4oTUAIun6Z8xw6mKQcAAHB3FCeghFw+Tfmnn0qPP246EQAAAIqK4gSUoIgI6aOP7PU332SacgAAAHdFcQJK2MCBTFMOAADg7ihOQClgmnIAAAD3RnECSoHDIc2aJd16qz3DXo8e0pEjplMBAACgoChOQCkpV86eprx5c7s09erFNOUAAADuguIElKIqVZimHAAAwB1RnIBS1qCBtHKl5O9vT1P+j3+YTgQAAID8UJwAA6677tI05W+9JU2fbjYPAAAA8kZxAgy56y5p6lR7fcwYewgfAAAAXBPFCTDoqaekv/3t0jTlcXGmEwEAACAnFCfAIIdDmjnTnqY8NVXq2ZNpygEAAFwRxQkwLHOa8hYtmKYcAADAVVGcABdQpYr02WdSrVr2NOVDhjBNOQAAgCuhOAEu4vJpyj/7TBo3znQiAAAAZKI4AS6kY0dp/nx7/e23pXfeMZsHAAAANooT4GIGDJBeftleHzvWvvoEAAAAs1yiOM2YMUOhoaHy9/dXx44dtXnz5gIdt3DhQjkcDvXp06dkAwKl7Mknpfvus6cpHzzYvu8JAAAA5hgvTosWLdK4ceM0efJkbd++XWFhYeratauOHz+e53GHDh3S448/rs6dO5dSUqD0OBzSu+9KkZGXpin/7TfTqQAAAMou48Xp9ddf19///neNHDlSLVq00KxZs1ShQgV98MEHuR6Tnp6uYcOG6fnnn1fDhg1LMS1QesqVk/7zH3ua8qNHmaYcAADAJKPF6fz589q2bZsiIyOd27y8vBQZGalNmzbletyUKVNUq1Yt/e1vf8v3Z6SlpSk5OTnLAriLKlWkVavsacrj4uxhe0xTDgAAUPqMFqeTJ08qPT1dQUFBWbYHBQUpISEhx2O++eYbvf/++5ozZ06BfsbUqVMVGBjoXEJCQq44N1CaQkMvTVO+apX02GOmEwEAAJQ9xofqFUZKSoruuecezZkzRzVq1CjQMRMmTFBSUpJziY+PL+GUQPHr2FH6+GN7/Z137KnKAQAAUHp8TP7wGjVqyNvbW4mJiVm2JyYmqnbt2tn2/+WXX3To0CH16tXLuS0jI0OS5OPjo71796pRo0ZZjvHz85Ofn18JpAdKV//+0rRp0lNP2VedGjSw73sCAABAyTN6xcnX11ft2rVTbGysc1tGRoZiY2MVERGRbf9mzZpp165diouLcy533nmnunTpori4OIbhweM98YT097/b05QPGcI05QAAAKXF6BUnSRo3bpxGjBih9u3bq0OHDnrzzTeVmpqqkSNHSpKGDx+uunXraurUqfL391erVq2yHF+lShVJyrYd8EQOhzRjhnTokBQTY09T/v33Ur16ppMBAAB4NuPFadCgQTpx4oQmTZqkhIQEtWnTRmvXrnVOGHH48GF5ebnVrVhAicqcpvz666Uff7TL09dfS5Urm04GAADguRyWZVmmQ5Sm5ORkBQYGKikpSQEBAabjAEV26JA9acTx41L37tKKFZKP8T+FAAAAuI/CdAMu5QBuKjRU+vRTe5ry1auZphwAAKAkUZwAN9ahgz1NucMhTZ/ONOUAAAAlheIEuLnMacolaexY+yoUAAAAihfFCfAAjz8u3X+/ZFnS4MHS9u2mEwEAAHgWihPgATKH6t12m3T2rD3TXny86VQAAACeg+IEeIjMacpbtpSOHbPLU0qK6VQAAACegeIEeJDAQGnVKikoSPrf/+xhexcvmk4FAADg/ihOgIepX19auVIqX96epnzsWPveJwAAABQdxQnwQB06SFFR9r1PM2bY05Snp0sbNkjR0fbH9HTTKQEAANwHxQnwUH37Sq+8Yq+PHWsP3+vSRRo61P4YGiotXWoyIQAAgPugOAEe7B//sGfak6Tff8/6tSNHpAEDKE8AAAAFQXECPFhGhvTjjzl/LfO+p7FjGbYHAACQH4oT4MG+/tq+spQby7Kf9/T116WXCQAAwB1RnAAPduxY8e4HAABQVlGcAA8WHFyw/SZOlN56Szp5smTzAAAAuCuKE+DBOneW6tWzpyXPyy+/2Pc61akj9etnPwfqwoVSiQgAAOAWKE6AB/P2tq8kSdnLk8NhL3PnStOnS+3b22Vp2TKpd2+pbl3psceknTtLPTYAAIDLoTgBHq5fP2nJErsIXa5ePXv7iBHS6NHSli3Srl3S44/bz3w6cUJ6802pTRspPNwuYCdOmHgFAAAA5jksK3NS4rIhOTlZgYGBSkpKUkBAgOk4QKlJT7dnzzt2zL73qXNn+4pUTi5elNats69GrVwpnT9vb/fxkXr2lO69V+reXSpXrrTSAwAAFL/CdAOKE4A8/f67tHChXaK2br20vWZNadgwu0SFhZlKBwAAUHQUpzxQnICi271bmjdPmj9fSky8tL1NG7tADR1qFyoAAAB3QHHKA8UJuHIM5QMAAJ6A4pQHihNQvBjKBwAA3BXFKQ8UJ6DkMJQPAAC4E4pTHihOQMm7eFFav96+CrViBUP5AACAa6I45YHiBJSuU6cuDeXbsuXSdobyAQAA0yhOeaA4Aeb88MOloXwJCZe2M5QPAACYQHHKA8UJMC+voXw9elwayufrazIlAADwdBSnPFCcANeS21C+GjUuDeVr08ZQOAAA4NEoTnmgOAGuK7ehfGFhl4by1aplLB4AAPAwFKc8UJwA18dQPgAAUBooTnmgOAHuhaF8AACgpFCc8kBxAtwXQ/kAAEBxKkw38CqlTHmaMWOGQkND5e/vr44dO2rz5s257jtnzhx17txZVatWVdWqVRUZGZnn/gA8R8uW0iuvSPHx0qpV0l132cP1du6UHntMqltX6tNHWr780vA+AACA4mC8OC1atEjjxo3T5MmTtX37doWFhalr1646fvx4jvtv2LBBQ4YM0ZdffqlNmzYpJCREt99+u44cOVLKyQGY4uNj3+O0eLF07Jg0Y4Z07bX2vVErVkh9+9olauxYKS7OdFoAAOAJjA/V69ixo6699lpNnz5dkpSRkaGQkBA98sgjGj9+fL7Hp6enq2rVqpo+fbqGDx+e7/4M1QM8148/2kP5PvqIoXwAACB/bjNU7/z589q2bZsiIyOd27y8vBQZGalNmzYV6HucPXtWFy5cULVq1XL8elpampKTk7MsADxTixbStGn2UL7Vq6WBAxnKBwAAiofR4nTy5Emlp6crKCgoy/agoCAlXP7n4jw89dRTqlOnTpbydbmpU6cqMDDQuYSEhFxxbgCuzcdH6tZNWrTIHsr37rtShw4M5QMAAEVn/B6nK/Hyyy9r4cKFWrZsmfz9/XPcZ8KECUpKSnIu8fHxpZwSgEnVqkkPPih9/709K9+TT0q1a0snT0pvvSWFh9tD+d54Q8rp1sr0dGnDBik62v6Ynl7arwAAALgCo8WpRo0a8vb2VmJiYpbtiYmJql27dp7H/utf/9LLL7+s9evXq3Xr1rnu5+fnp4CAgCwLgLIpt6F8//ufNG6cfRWqd+9LQ/mWLpVCQ6UuXez7o7p0sT9futTwCwEAAKXOaHHy9fVVu3btFBsb69yWkZGh2NhYRURE5HrcK6+8ohdeeEFr165V+/btSyMqAA+S11C+lSvtoXw1akj9+0u//Zb12CNHpAEDKE8AAJQ1xofqjRs3TnPmzNG8efO0Z88ePfjgg0pNTdXIkSMlScOHD9eECROc+0+bNk0TJ07UBx98oNDQUCUkJCghIUFnzpwx9RIAuLHchvKlpOS8f+Y8pGPHMmwPAICyxMd0gEGDBunEiROaNGmSEhIS1KZNG61du9Y5YcThw4fl5XWp382cOVPnz5/XgAEDsnyfyZMn67nnnivN6AA8TOZQvttvl3KZb0aSXZ7i46Wrr5aaN5dCQrIv9epJudx6CQAA3JDx5ziVNp7jBCA/0dH2PU1XqmbNnEtV5lKnjlSu3JX/HAAAUDSF6QbGrzgBgKsJDi7Yfi+/LFWtal99+uty7px04oS9bN+e8/FeXvawwLzKVVCQ5O1dfK8NAAAUDVecAOAv0tPt2fOOHLl0T9PlHA57KN7BgzmXGsuSTp3KuVBlLr/9Jl24kH8WHx97tr+8ylWNGnYmAABQOFxxAoAr4O1tP+NpwAC7kFxenjILyptv5n4lyOGQqle3lzZtct4nI8N+blRe5eroUXumv19/tZfc+PnZRS6vclWlCuUKAIArwRUnAMjF0qXSmDFZpyQPCbFLU79+Jf/zL160p0vP7YpVfLyUkFCw71WxYt7FKiREqlSpePOnp0tff22/huBgqXNnhh0CAFxLYboBxQkA8uDqb/7Pn7eHFOZ15er33wv2vapUybtYFWamwJxKZ7169pW80iidAFDSXP3fB1flaueN4pQHihOAsubs2UtXqHJbkpML9r1ymykwc6hg3brSp5/awxz/+q9L5lDBJUsoT3lxtTcV7oLzVnScu8Ljj0NF44rnjeKUB4oTAGSXnJx1CGBOy9mzBfteXl72PVy5qVZNmj5dqlDBvoKV0+Lnl3W9rNyf5YpvKtwB563oOHeFt3QpfxwqClc9bxSnPFCcAKDwLEv644/8Zwo8f75kfv7lRaq0l9Iqbq76psLVcd6KjnNXeJmzrl5eNC+X16yrlmUvGRlml/T00v+ZFy/a9wfnNrohv9lqSxLFKQ8UJwAoGRkZ0uzZ0oMP5r9vixZSYKD9vKvcFlf61ymv4lYcpa5cOal/f3umxZw4HPYDk3futKeoz9z216Uo293ZlbyJLQmZb4rd4eOFC/aDvk+ezP31VK8uzZhhn8fifCPuzvufPZv7/08v5+ub/by50n/TXNWXX0o331y6P5PpyAEApc7LS2rWrGD7zpiR9z+OmW/s8ipWmUtaWsH2K8zy559Z3+SkpdlLUtIVnaIisyx7EpAaNUrm+xdHCSutYy7fnnn/Xm4sy74aWr++XVBLsoh44pvi33+XBg82ncI9FcfVdy+v0lm8vUv+Z/zyi/T55/m/5mPHrvy8lSSKEwCg2HTubP+FP7+HB3funPf3cTjsv9j6+komBgdYlj20pLgLWW6F78SJgv0VuyRfrye+8c905IjpBPnLLINeXsX/Madtycl5l85MzZpJtWsX/5vvkt6/pH7Gjh3SAw/kf96io6VOnYpeNDzhivDlNmwoWHEKDi7xKFeEoXoAgGKVed+ElPPDg7lvIrsNG6QuXfLfb/166aabLhWdvy5S2frajh3SY4/lf97efltq167kiklxfCxtBf2dMzF0ypVlDg/N749DJu7VcWWufN64xykPFCcAKHmmHx7sblz5TYUr47wVHeeu6PjjUNG46nkrTDfwKqVMAIAypF8/6dAh+6/VCxbYHw8e5M1Ebry97emfpexXHzI/f/NN3sD+Feet6Dh3Rdevn/0mv27drNvr1aM05cUTzhtXnAAAcBFcqSsazlvRce6KjgcHF42rnTeG6uWB4gQAcGWu9qbCXXDeio5zh7KM4pQHihMAAAAAiXucAAAAAKBYUZwAAAAAIB8UJwAAAADIB8UJAAAAAPJBcQIAAACAfFCcAAAAACAfFCcAAAAAyAfFCQAAAADyQXECAAAAgHxQnAAAAAAgHz6mA5Q2y7IkScnJyYaTAAAAADApsxNkdoS8lLnilJKSIkkKCQkxnAQAAACAK0hJSVFgYGCe+zisgtQrD5KRkaGjR4+qcuXKcjgcpuMoOTlZISEhio+PV0BAgOk4boPzVjSct6LhvBUd565oOG9Fw3krOs5d0XDeisaVzptlWUpJSVGdOnXk5ZX3XUxl7oqTl5eX6tWrZzpGNgEBAcZ/cdwR561oOG9Fw3krOs5d0XDeiobzVnScu6LhvBWNq5y3/K40ZWJyCAAAAADIB8UJAAAAAPJBcTLMz89PkydPlp+fn+koboXzVjSct6LhvBUd565oOG9Fw3krOs5d0XDeisZdz1uZmxwCAAAAAAqLK04AAAAAkA+KEwAAAADkg+IEAAAAAPmgOAEAAABAPihOhnz11Vfq1auX6tSpI4fDoeXLl5uO5BamTp2qa6+9VpUrV1atWrXUp08f7d2713Qslzdz5ky1bt3a+aC5iIgIrVmzxnQst/Pyyy/L4XBo7NixpqO4tOeee04OhyPL0qxZM9Ox3MaRI0d09913q3r16ipfvryuueYabd261XQslxYaGprtd87hcGj06NGmo7m09PR0TZw4UQ0aNFD58uXVqFEjvfDCC2LesPylpKRo7Nixql+/vsqXL6/rr79eW7ZsMR3L5eT3fteyLE2aNEnBwcEqX768IiMjtW/fPjNhC4DiZEhqaqrCwsI0Y8YM01HcysaNGzV69Gj997//VUxMjC5cuKDbb79dqamppqO5tHr16unll1/Wtm3btHXrVt1yyy3q3bu3fvjhB9PR3MaWLVv03nvvqXXr1qajuIWWLVvq2LFjzuWbb74xHckt/PHHH+rUqZPKlSunNWvW6Mcff9Rrr72mqlWrmo7m0rZs2ZLl9y0mJkaSdNdddxlO5tqmTZummTNnavr06dqzZ4+mTZumV155Re+8847paC7vvvvuU0xMjObPn69du3bp9ttvV2RkpI4cOWI6mkvJ7/3uK6+8orfffluzZs3S999/r4oVK6pr1646d+5cKSctIAvGSbKWLVtmOoZbOn78uCXJ2rhxo+kobqdq1arWv//9b9Mx3EJKSop19dVXWzExMdZNN91kjRkzxnQklzZ58mQrLCzMdAy39NRTT1k33HCD6Rhub8yYMVajRo2sjIwM01FcWo8ePaxRo0Zl2davXz9r2LBhhhK5h7Nnz1re3t7WZ599lmV727ZtrWeeecZQKtf31/e7GRkZVu3ata1XX33Vue306dOWn5+fFR0dbSBh/rjiBLeWlJQkSapWrZrhJO4jPT1dCxcuVGpqqiIiIkzHcQujR49Wjx49FBkZaTqK29i3b5/q1Kmjhg0batiwYTp8+LDpSG5h5cqVat++ve666y7VqlVL4eHhmjNnjulYbuX8+fP6+OOPNWrUKDkcDtNxXNr111+v2NhY/fzzz5KknTt36ptvvlG3bt0MJ3NtFy9eVHp6uvz9/bNsL1++PFfXC+HgwYNKSEjI8m9rYGCgOnbsqE2bNhlMljsf0wGAosrIyNDYsWPVqVMntWrVynQcl7dr1y5FRETo3LlzqlSpkpYtW6YWLVqYjuXyFi5cqO3btzN2vRA6duyouXPnqmnTpjp27Jief/55de7cWbt371blypVNx3NpBw4c0MyZMzVu3Dg9/fTT2rJlix599FH5+vpqxIgRpuO5heXLl+v06dO69957TUdxeePHj1dycrKaNWsmb29vpaen68UXX9SwYcNMR3NplStXVkREhF544QU1b95cQUFBio6O1qZNm9S4cWPT8dxGQkKCJCkoKCjL9qCgIOfXXA3FCW5r9OjR2r17N3/dKaCmTZsqLi5OSUlJWrJkiUaMGKGNGzdSnvIQHx+vMWPGKCYmJttfFpG7y/9a3bp1a3Xs2FH169fX4sWL9be//c1gMteXkZGh9u3b66WXXpIkhYeHa/fu3Zo1axbFqYDef/99devWTXXq1DEdxeUtXrxYUVFRWrBggVq2bKm4uDiNHTtWderU4fctH/Pnz9eoUaNUt25deXt7q23bthoyZIi2bdtmOhpKEEP14JYefvhhffbZZ/ryyy9Vr14903Hcgq+vrxo3bqx27dpp6tSpCgsL01tvvWU6lkvbtm2bjh8/rrZt28rHx0c+Pj7auHGj3n77bfn4+Cg9Pd10RLdQpUoVNWnSRPv37zcdxeUFBwdn+2NG8+bNGepYQL/++qs+//xz3XfffaajuIUnnnhC48eP1+DBg3XNNdfonnvu0WOPPaapU6eajubyGjVqpI0bN+rMmTOKj4/X5s2bdeHCBTVs2NB0NLdRu3ZtSVJiYmKW7YmJic6vuRqKE9yKZVl6+OGHtWzZMn3xxRdq0KCB6UhuKyMjQ2lpaaZjuLRbb71Vu3btUlxcnHNp3769hg0bpri4OHl7e5uO6BbOnDmjX375RcHBwaajuLxOnTple8TCzz//rPr16xtK5F4+/PBD1apVSz169DAdxS2cPXtWXl5Z3wp6e3srIyPDUCL3U7FiRQUHB+uPP/7QunXr1Lt3b9OR3EaDBg1Uu3ZtxcbGOrclJyfr+++/d9l7sBmqZ8iZM2ey/PX14MGDiouLU7Vq1XTVVVcZTObaRo8erQULFmjFihWqXLmycwxsYGCgypcvbzid65owYYK6deumq666SikpKVqwYIE2bNigdevWmY7m0ipXrpzt/rmKFSuqevXq3FeXh8cff1y9evVS/fr1dfToUU2ePFne3t4aMmSI6Wgu77HHHtP111+vl156SQMHDtTmzZs1e/ZszZ4923Q0l5eRkaEPP/xQI0aMkI8Pb28KolevXnrxxRd11VVXqWXLltqxY4def/11jRo1ynQ0l7du3TpZlqWmTZtq//79euKJJ9SsWTONHDnSdDSXkt/73bFjx+qf//ynrr76ajVo0EATJ05UnTp11KdPH3Oh82J6Wr+y6ssvv7QkZVtGjBhhOppLy+mcSbI+/PBD09Fc2qhRo6z69etbvr6+Vs2aNa1bb73VWr9+velYbonpyPM3aNAgKzg42PL19bXq1q1rDRo0yNq/f7/pWG7j008/tVq1amX5+flZzZo1s2bPnm06kltYt26dJcnau3ev6ShuIzk52RozZox11VVXWf7+/lbDhg2tZ555xkpLSzMdzeUtWrTIatiwoeXr62vVrl3bGj16tHX69GnTsVxOfu93MzIyrIkTJ1pBQUGWn5+fdeutt7r0/4cdlsXjoQEAAAAgL9zjBAAAAAD5oDgBAAAAQD4oTgAAAACQD4oTAAAAAOSD4gQAAAAA+aA4AQAAAEA+KE4AAAAAkA+KEwAAAADkg+IEAEAhOBwOLV++3HQMAEApozgBANzGvffeK4fDkW254447TEcDAHg4H9MBAAAojDvuuEMffvhhlm1+fn6G0gAAygquOAEA3Iqfn59q166dZalataokexjdzJkz1a1bN5UvX14NGzbUkiVLshy/a9cu3XLLLSpfvryqV6+u+++/X2fOnMmyzwcffKCWLVvKz89PwcHBevjhh7N8/eTJk+rbt68qVKigq6++WitXrizZFw0AMI7iBADwKBMnTlT//v21c+dODRs2TIMHD9aePXskSampqeratauqVq2qLVu26D//+Y8+//zzLMVo5syZGj16tO6//37t2rVLK1euVOPGjbP8jOeff14DBw7U//73P3Xv3l3Dhg3TqVOnSvV1AgBKl8OyLMt0CAAACuLee+/Vxx9/LH9//yzbn376aT399NNyOBx64IEHNHPmTOfXrrvuOrVt21bvvvuu5syZo6eeekrx8fGqWLGiJGn16tXq1auXjh49qqCgINWtW1cjR47UP//5zxwzOBwOPfvss3rhhRck2WWsUqVKWrNmDfdaAYAH4x4nAIBb6dKlS5ZiJEnVqlVzrkdERGT5WkREhOLi4iRJe/bsUVhYmLM0SVKnTp2UkZGhvXv3yuFw6OjRo7r11lvzzNC6dWvnesWKFRUQEKDjx48X9SUBANwAxQkA4FYqVqyYbehccSlfvnyB9itXrlyWzx0OhzIyMkoiEgDARXCPEwDAo/z3v//N9nnz5s0lSc2bN9fOnTuVmprq/Pq3334rLy8vNW3aVJUrV1ZoaKhiY2NLNTMAwPVxxQkA4FbS0tKUkJCQZZuPj49q1KghSfrPf/6j9u3b64YbblBUVJQ2b96s999/X5I0bNgwTZ48WSNGjNBzzz2nEydO6JFHHtE999yjoKAgSdJzzz2nBx54QLVq1VK3bt2UkpKib7/9Vo888kjpvlAAgEuhOAEA3MratWsVHBycZVvTpk31008/SbJnvFu4cKEeeughBQcHKzo6Wi1atJAkVahQQevWrdOYMWN07bXXqkKFCurfv79ef/115/caMWKEzp07pzfeeEOPP/64atSooQEDBpTeCwQAuCRm1QMAeAyHw6Fly5apT58+pqMAADwM9zgBAAAAQD4oTgAAAACQD+5xAgB4DEafAwBKClecAAAAACAfFCcAAAAAyAfFCQAAAADyQXECAAAAgHxQnAAAAAAgHxQnAAAAAMgHxQkAAAAA8kFxAgAAAIB8/D9+bdOrEm6qZgAAAABJRU5ErkJggg==",
+      "image/png": "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",
       "text/plain": [
        "<Figure size 1000x500 with 1 Axes>"
       ]
@@ -168,6 +20168,13 @@
     "plt.xticks(epochs_list)\n",
     "plt.show()"
    ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
   }
  ],
  "metadata": {
diff --git a/norch/__init__.py b/norch/__init__.py
index ca7be0b..3eb0b7f 100644
--- a/norch/__init__.py
+++ b/norch/__init__.py
@@ -2,6 +2,7 @@ from norch.tensor import Tensor
 from .nn import *
 from .optim import *
 from .utils import *
+from .datasets import *
 
 __version__ = "0.0.1"
 __author__ = 'Lucas de Lima Nogueira'
diff --git a/norch/__pycache__/__init__.cpython-38.pyc b/norch/__pycache__/__init__.cpython-38.pyc
index d172276..8202d1d 100644
Binary files a/norch/__pycache__/__init__.cpython-38.pyc and b/norch/__pycache__/__init__.cpython-38.pyc differ
diff --git a/norch/__pycache__/tensor.cpython-38.pyc b/norch/__pycache__/tensor.cpython-38.pyc
index 5a33b4a..bcc1fe0 100644
Binary files a/norch/__pycache__/tensor.cpython-38.pyc and b/norch/__pycache__/tensor.cpython-38.pyc differ
diff --git a/norch/autograd/__pycache__/functions.cpython-38.pyc b/norch/autograd/__pycache__/functions.cpython-38.pyc
index d48733e..71b5662 100644
Binary files a/norch/autograd/__pycache__/functions.cpython-38.pyc and b/norch/autograd/__pycache__/functions.cpython-38.pyc differ
diff --git a/norch/autograd/functions.py b/norch/autograd/functions.py
index 8ce8625..c9a393e 100644
--- a/norch/autograd/functions.py
+++ b/norch/autograd/functions.py
@@ -1,4 +1,5 @@
 import math
+import norch
 
 class AddBackward:
     def __init__(self, x, y):
@@ -15,6 +16,7 @@ class AddBroadcastedBackward:
         x, y = self.input
         grad_x = self._reshape_gradient(gradient, x.shape)
         grad_y = self._reshape_gradient(gradient, y.shape)
+
         return [grad_x, grad_y]
     
     def _reshape_gradient(self, gradient, shape):
@@ -25,11 +27,17 @@ class AddBroadcastedBackward:
         # Sum along axes where the target shape dimension is 1
         for i in range(len(shape)):
             if shape[i] == 1:
-                gradient = gradient.sum(axis=i)
+                gradient = gradient.sum(axis=i, keepdim=True)
 
         return gradient
     
+class SubBackward:
+    def __init__(self, x, y):
+        self.input = [x, y]
 
+    def backward(self, gradient):
+        return [gradient, -gradient]
+    
 class SubBroadcastedBackward:
     def __init__(self, x, y):
         self.input = [x, y]
@@ -49,15 +57,7 @@ class SubBroadcastedBackward:
         for i in range(len(shape)):
             if shape[i] == 1:
                 gradient = gradient.sum(axis=i)
-
         return gradient
-
-class SubBackward:
-    def __init__(self, x, y):
-        self.input = [x, y]
-
-    def backward(self, gradient):
-        return [gradient, -gradient]
     
 class ScalarMulBackward:
     def __init__(self, x, scalar):
@@ -82,7 +82,13 @@ class MatmulBackward:
 
     def backward(self, gradient):
         x, y = self.input
-        return [gradient @ y.transpose(-1,-2), x.transpose(-1,-2) @ gradient]
+        
+        if x.ndim != y.ndim: # broadcasted case
+            aux = (gradient @ y.transpose(-1,-2))
+            aux_sum = aux.sum(axis=0)
+            return [aux_sum, x.transpose(-1,-2) @ gradient]
+        else:
+            return [gradient @ y.transpose(-1,-2), x.transpose(-1,-2) @ gradient]
         
 """class PowBackward:
     def __init__(self, x, power):
@@ -121,12 +127,29 @@ class LogBackward:
         return [grad_input]
        
 class SumBackward:
-    def __init__(self, x):
+    def __init__(self, x, axis=None, keepdim=False):
         self.input = [x]
+        self.axis = axis
+        self.keepdim = keepdim
 
     def backward(self, gradient):
-        # Since sum reduces a tensor to a scalar, gradient is broadcasted to match the original shape.
-        return [float(gradient.tensor.contents.data[0]) * self.input[0].ones_like()]
+        input_shape = self.input[0].shape.copy()
+        if self.axis == -1:
+            # If axis is None, sum reduces the tensor to a scalar.
+            grad_output = float(gradient.tensor.contents.data[0]) * self.input[0].ones_like()
+        else:
+
+            if self.keepdim:
+                input_shape = input_shape[:self.axis] + [1] + input_shape[self.axis+1:]
+            else:
+                input_shape = input_shape[:self.axis] + input_shape[self.axis+1:]
+
+            # Broadcast the gradient to the input shape along the specified axis.
+            grad_output_shape = list(input_shape)
+            grad_output = gradient.reshape(grad_output_shape)
+            grad_output = grad_output + self.input[0].zeros_like()
+        
+        return [grad_output]
     
 class ReshapeBackward:
     def __init__(self, x):
@@ -177,5 +200,95 @@ class CosBackward:
     def backward(self, gradient):
         x = self.input[0]
         return [-gradient * x.sin()]
+    
+class MaxBackward:
+    def __init__(self, x, axis=None, keepdim=False):
+        self.input = [x]
+        self.axis = axis
+        self.keepdim = keepdim
+
+    def backward(self, gradient):
+        input_shape = self.input[0].shape.copy()
+        if self.axis == -1:
+            max_value = self.input[0].max()
+            mask = self.input[0].equal(max_value)
+
+            grad_output = float(gradient.tensor.contents.data[0]) * self.input[0].ones_like()
+
+            grad_output = (grad_output * mask) / mask.sum().tensor.contents.data[0]
+
+        else:
+
+            if self.keepdim:
+                input_shape = input_shape[:self.axis] + [1] + input_shape[self.axis+1:]
+            else:
+                input_shape = input_shape[:self.axis] + input_shape[self.axis+1:]
+
+            # Broadcast the gradient to the input shape along the specified axis.
+            grad_output_shape = list(input_shape)
+                            
+            grad_output = gradient.reshape(grad_output_shape)
+            grad_output = grad_output + self.input[0].zeros_like()
+            max_values = self.input[0].max(axis=self.axis, keepdim=True)
+            mask = self.input[0].equal(max_values)
+
+            grad_output = (grad_output * mask)
+        
+        return [grad_output]
+
+    
+class MinBackward:
+    def __init__(self, x, axis=None, keepdim=False):
+        self.input = [x]
+        self.axis = axis
+        self.keepdim = keepdim
+
+    def backward(self, gradient):
+        input_shape = self.input[0].shape.copy()
+        if self.axis == -1:
+            min_value = self.input[0].min()
+            mask = self.input[0].equal(min_value)
+
+            grad_output = float(gradient.tensor.contents.data[0]) * self.input[0].ones_like()
+
+            grad_output = (grad_output * mask) / mask.sum().tensor.contents.data[0]
+
+        else:
+            if self.keepdim:
+                input_shape = input_shape[:self.axis] + [1] + input_shape[self.axis+1:]
+            else:
+                input_shape = input_shape[:self.axis] + input_shape[self.axis+1:]
+
+            # Broadcast the gradient to the input shape along the specified axis.
+            grad_output_shape = list(input_shape)
+            grad_output = gradient.reshape(grad_output_shape)
+            grad_output = grad_output + self.input[0].zeros_like()
+            max_values = self.input[0].min(axis=self.axis, keepdim=True)
+            mask = self.input[0].equal(max_values)
+
+            grad_output = (grad_output * mask)
+
+        return [grad_output]
+
+
+class CrossEntropyLossBackward:
+    def __init__(self, logits, targets):
+        self.input = [logits, targets]
+    
+    def backward(self, gradient):
+        logits, targets = self.input
+
+        if logits.ndim == 1:
+            softmax = norch.softmax(logits, dim=0)
+            grad_logits = (softmax - targets)
+                
+        elif logits.ndim == 2:
+            # batched 
+                batch_size = logits.shape[0]
+                softmax = norch.softmax(logits, dim=1)
+
+                grad_logits = (softmax - targets) / batch_size
+
+        return [grad_logits, None]  # targets do not have a gradient
 
 
diff --git a/norch/csrc/cpu.cpp b/norch/csrc/cpu.cpp
index bde7b0c..9584c23 100644
--- a/norch/csrc/cpu.cpp
+++ b/norch/csrc/cpu.cpp
@@ -11,7 +11,7 @@ void add_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data) {
     }
 }
 
-void add_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape) {
+void add_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size) {
     int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
 
     // Calculate strides for broadcasting
@@ -28,12 +28,12 @@ void add_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_
         int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - max_ndim + i] : 1;
         strides1[i] = dim1 == broadcasted_shape[i] ? stride1 : 0;
         strides2[i] = dim2 == broadcasted_shape[i] ? stride2 : 0;
-        stride1 *= broadcasted_shape[i];
-        stride2 *= broadcasted_shape[i];
+        stride1 *= (dim1 == broadcasted_shape[i]) ? dim1 : 1;
+        stride2 *= (dim2 == broadcasted_shape[i]) ? dim2 : 1;
     }
 
     // Perform element-wise addition with broadcasting
-    for (int i = 0; i < tensor1->size; i++) {
+    for (int i = 0; i < broadcasted_size; i++) {
         int index1 = 0, index2 = 0;
         int linear_index = i;
         for (int j = max_ndim - 1; j >= 0; j--) {
@@ -58,7 +58,7 @@ void sub_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data) {
     }
 }
 
-void sub_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape) {
+void sub_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size) {
     int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
 
     // Calculate strides for broadcasting
@@ -75,12 +75,12 @@ void sub_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_
         int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - max_ndim + i] : 1;
         strides1[i] = dim1 == broadcasted_shape[i] ? stride1 : 0;
         strides2[i] = dim2 == broadcasted_shape[i] ? stride2 : 0;
-        stride1 *= broadcasted_shape[i];
-        stride2 *= broadcasted_shape[i];
+        stride1 *= (dim1 == broadcasted_shape[i]) ? dim1 : 1;
+        stride2 *= (dim2 == broadcasted_shape[i]) ? dim2 : 1;
     }
 
     // Perform element-wise addition with broadcasting
-    for (int i = 0; i < tensor1->size; i++) {
+    for (int i = 0; i < broadcasted_size; i++) {
         int index1 = 0, index2 = 0;
         int linear_index = i;
         for (int j = max_ndim - 1; j >= 0; j--) {
@@ -205,7 +205,7 @@ void log_tensor_cpu(Tensor* tensor, float* result_data) {
     }
 }
 
-void sum_tensor_cpu(Tensor* tensor, float* result_data, int axis) {
+void sum_tensor_cpu(Tensor* tensor, float* result_data, int size, int* result_shape, int axis) {
     if (axis == -1) {
         // Sum over all elements
         float sum = 0.0;
@@ -219,26 +219,14 @@ void sum_tensor_cpu(Tensor* tensor, float* result_data, int axis) {
             return;
         }
         
-        int result_shape[tensor->ndim - 1];
-        int result_size = 1;
         int axis_stride = tensor->strides[axis];
 
-        int idx = 0;
-        for (int i = 0; i < tensor->ndim; i++) {
-            if (i != axis) {
-                result_shape[idx++] = tensor->shape[i];
-                result_size *= tensor->shape[i];
-            }
-        }
-
-        memset(result_data, 0, result_size * sizeof(float));
-
         for (int i = 0; i < tensor->shape[axis]; i++) {
-            for (int j = 0; j < result_size; j++) {
+            for (int j = 0; j < size; j++) {
                 int index = 0;
                 int remainder = j;
                 for (int k = tensor->ndim - 2; k >= 0; k--) {
-                    index += (remainder % result_shape[k]) * tensor->strides[k < axis ? k : k + 1];
+                    index += (remainder % result_shape[k]) * tensor->strides[k < axis ? k : k + 1];     
                     remainder /= result_shape[k];
                 }
                 result_data[j] += tensor->data[index + i * axis_stride];
@@ -247,8 +235,115 @@ void sum_tensor_cpu(Tensor* tensor, float* result_data, int axis) {
     }
 }
 
+void max_tensor_cpu(Tensor* tensor, float* result_data, int size, int* result_shape, int axis) {
+    if (axis == -1) {
+        float max_value = -INFINITY;
+        for (int i = 0; i < tensor->size; i++) {
+            max_value = fmax(max_value, tensor->data[i]);
+        }
+        *result_data = max_value;
+    } else {
+        for (int i = 0; i < size; i++) {
+            result_data[i] = -INFINITY;
+        }
+        if (axis < 0 || axis >= tensor->ndim) {
+            printf("Invalid axis");
+            return;
+        }
+        
+        int axis_stride = tensor->strides[axis];
 
+        for (int i = 0; i < tensor->shape[axis]; i++) {
+            for (int j = 0; j < size; j++) {
+                int index = 0;
+                int remainder = j;
+                for (int k = tensor->ndim - 2; k >= 0; k--) {
+                    index += (remainder % result_shape[k]) * tensor->strides[k < axis ? k : k + 1];     
+                    remainder /= result_shape[k];
+                }
+                result_data[j] = fmax(result_data[j], tensor->data[index + i * axis_stride]);
+            }
+        }
+    }
+}
 
+void min_tensor_cpu(Tensor* tensor, float* result_data, int size, int* result_shape, int axis) {
+    if (axis == -1) {
+        float min_value = INFINITY;
+        for (int i = 0; i < tensor->size; i++) {
+            min_value = fmin(min_value, tensor->data[i]);
+        }
+        *result_data = min_value;
+    } else {
+        for (int i = 0; i < size; i++) {
+            result_data[i] = INFINITY;
+        }
+        if (axis < 0 || axis >= tensor->ndim) {
+            printf("Invalid axis");
+            return;
+        }
+        
+        int axis_stride = tensor->strides[axis];
+
+        for (int i = 0; i < tensor->shape[axis]; i++) {
+            for (int j = 0; j < size; j++) {
+                int index = 0;
+                int remainder = j;
+                for (int k = tensor->ndim - 2; k >= 0; k--) {
+                    index += (remainder % result_shape[k]) * tensor->strides[k < axis ? k : k + 1];     
+                    remainder /= result_shape[k];
+                }
+                result_data[j] = fmin(result_data[j], tensor->data[index + i * axis_stride]);
+            }
+        }
+    }
+}
+
+void equal_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data) {
+    
+    for (int i = 0; i < tensor1->size; i++) {
+        result_data[i] = (tensor1->data[i] == tensor2->data[i]) ? 1.0f : 0.0f;
+    }
+}
+
+void equal_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size) {
+    int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
+
+    // Calculate strides for broadcasting
+    int* strides1 = (int*)malloc(max_ndim * sizeof(int));
+    int* strides2 = (int*)malloc(max_ndim * sizeof(int));
+    if (strides1 == NULL || strides2 == NULL) {
+        fprintf(stderr, "Memory allocation failed\n");
+        exit(1);
+    }
+
+    int stride1 = 1, stride2 = 1;
+    for (int i = max_ndim - 1; i >= 0; i--) {
+        int dim1 = i < tensor1->ndim ? tensor1->shape[tensor1->ndim - max_ndim + i] : 1;
+        int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - max_ndim + i] : 1;
+        strides1[i] = dim1 == broadcasted_shape[i] ? stride1 : 0;
+        strides2[i] = dim2 == broadcasted_shape[i] ? stride2 : 0;
+        stride1 *= (dim1 == broadcasted_shape[i]) ? dim1 : 1;
+        stride2 *= (dim2 == broadcasted_shape[i]) ? dim2 : 1;
+    }
+
+    // Perform element-wise equal with broadcasting
+    for (int i = 0; i < broadcasted_size; i++) {
+        int index1 = 0, index2 = 0;
+        int linear_index = i;
+        for (int j = max_ndim - 1; j >= 0; j--) {
+            int pos = linear_index % broadcasted_shape[j];
+            linear_index /= broadcasted_shape[j];
+            if (strides1[j] != 0) index1 += pos * strides1[j];
+            if (strides2[j] != 0) index2 += pos * strides2[j];
+        }
+        result_data[i] = (tensor1->data[index1] == tensor2->data[index2]) ? 1.0f : 0.0f;
+    }
+
+    // Free strides
+    free(strides1);
+    free(strides2);
+}
 
 
 void ones_like_tensor_cpu(Tensor* tensor, float* result_data) {
diff --git a/norch/csrc/cpu.h b/norch/csrc/cpu.h
index fad467c..83bd83f 100644
--- a/norch/csrc/cpu.h
+++ b/norch/csrc/cpu.h
@@ -4,10 +4,12 @@
 #include "tensor.h"
 
 void add_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data);
-void add_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape);
-void sum_tensor_cpu(Tensor* tensor, float* result_data, int axis);
+void add_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size);
+void sum_tensor_cpu(Tensor* tensor, float* result_data, int size, int* shape, int axis);
+void max_tensor_cpu(Tensor* tensor, float* result_data, int size, int* result_shape, int axis);
+void min_tensor_cpu(Tensor* tensor, float* result_data, int size, int* result_shape, int axis);
 void sub_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data);
-void sub_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape);
+void sub_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size);
 void elementwise_mul_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data);
 void scalar_div_tensor_cpu(float scalar, Tensor* tensor, float* result_data);
 void tensor_div_scalar_cpu(Tensor* tensor, float scalar, float* result_data);
@@ -19,6 +21,8 @@ void scalar_pow_tensor_cpu(float base, Tensor* tensor, float* result_data);
 void tensor_pow_scalar_cpu(Tensor* tensor, float exponent, float* result_data);
 void log_tensor_cpu(Tensor* tensor, float* result_data);
 void scalar_mul_tensor_cpu(Tensor* tensor, float scalar, float* result_data);
+void equal_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data);
+void equal_broadcasted_tensor_cpu(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size);
 void ones_like_tensor_cpu(Tensor* tensor, float* result_data);
 void zeros_like_tensor_cpu(Tensor* tensor, float* result_data);
 void transpose_1D_tensor_cpu(Tensor* tensor, float* result_data);
diff --git a/norch/csrc/cuda.cu b/norch/csrc/cuda.cu
index 9ebe6d3..7d056ab 100644
--- a/norch/csrc/cuda.cu
+++ b/norch/csrc/cuda.cu
@@ -71,7 +71,7 @@ __global__ void add_broadcasted_tensor_cuda_kernel(float* data1, float* data2, f
     result_data[i] = data1[index1] + data2[index2];
 }
 
-__host__ void add_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape) {
+__host__ void add_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size) {
     int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
 
     int* strides1 = (int*)malloc(max_ndim * sizeof(int));
@@ -87,8 +87,8 @@ __host__ void add_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, floa
         int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - max_ndim + i] : 1;
         strides1[i] = dim1 == broadcasted_shape[i] ? stride1 : 0;
         strides2[i] = dim2 == broadcasted_shape[i] ? stride2 : 0;
-        stride1 *= broadcasted_shape[i];
-        stride2 *= broadcasted_shape[i];
+        stride1 *= (dim1 == broadcasted_shape[i]) ? dim1 : 1;
+        stride2 *= (dim2 == broadcasted_shape[i]) ? dim2 : 1;
     }
     
     int* d_broadcasted_shape;
@@ -104,8 +104,8 @@ __host__ void add_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, floa
     cudaMalloc((void**)&d_strides2, max_ndim * sizeof(int));
     cudaMemcpy(d_strides2, strides2, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
 
-    int number_of_blocks = (tensor1->size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
-    add_broadcasted_tensor_cuda_kernel<<<number_of_blocks, THREADS_PER_BLOCK>>>(tensor1->data, tensor2->data, result_data, d_broadcasted_shape, d_strides1, d_strides2, max_ndim, tensor1->size);
+    int number_of_blocks = (broadcasted_size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
+    add_broadcasted_tensor_cuda_kernel<<<number_of_blocks, THREADS_PER_BLOCK>>>(tensor1->data, tensor2->data, result_data, d_broadcasted_shape, d_strides1, d_strides2, max_ndim, broadcasted_size);
 
     cudaError_t error = cudaGetLastError();
     if (error != cudaSuccess) {
@@ -187,6 +187,67 @@ __host__ void sub_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_da
     cudaDeviceSynchronize();
 }
 
+__global__ void sub_broadcasted_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int* broadcasted_shape, int* strides1, int*strides2, int max_ndim, int size) {
+    int i = blockIdx.x * blockDim.x + threadIdx.x;
+    if (i >= size) return;
+    
+    int index1 = 0, index2 = 0;
+    int linear_index = i;
+    for (int j = max_ndim - 1; j >= 0; j--) {
+        int pos = linear_index % broadcasted_shape[j];
+        linear_index /= broadcasted_shape[j];
+        if (strides1[j] != 0) index1 += pos * strides1[j];
+        if (strides2[j] != 0) index2 += pos * strides2[j];
+    }
+    result_data[i] = data1[index1] - data2[index2];
+}
+
+__host__ void sub_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size) {
+    int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
+
+    int* strides1 = (int*)malloc(max_ndim * sizeof(int));
+    int* strides2 = (int*)malloc(max_ndim * sizeof(int));
+    if (strides1 == NULL || strides2 == NULL) {
+        fprintf(stderr, "Memory allocation failed\n");
+        exit(1);
+    }
+
+    int stride1 = 1, stride2 = 1;
+    for (int i = max_ndim - 1; i >= 0; i--) {
+        int dim1 = i < tensor1->ndim ? tensor1->shape[tensor1->ndim - max_ndim + i] : 1;
+        int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - max_ndim + i] : 1;
+        strides1[i] = dim1 == broadcasted_shape[i] ? stride1 : 0;
+        strides2[i] = dim2 == broadcasted_shape[i] ? stride2 : 0;
+        stride1 *= (dim1 == broadcasted_shape[i]) ? dim1 : 1;
+        stride2 *= (dim2 == broadcasted_shape[i]) ? dim2 : 1;
+    }
+    
+    int* d_broadcasted_shape;
+    int* d_strides1;
+    int* d_strides2;
+
+    cudaMalloc((void**)&d_broadcasted_shape, max_ndim * sizeof(int));
+    cudaMemcpy(d_broadcasted_shape, broadcasted_shape, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
+
+    cudaMalloc((void**)&d_strides1, max_ndim * sizeof(int));
+    cudaMemcpy(d_strides1, strides1, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
+
+    cudaMalloc((void**)&d_strides2, max_ndim * sizeof(int));
+    cudaMemcpy(d_strides2, strides2, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
+
+    int number_of_blocks = (broadcasted_size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
+    sub_broadcasted_tensor_cuda_kernel<<<number_of_blocks, THREADS_PER_BLOCK>>>(tensor1->data, tensor2->data, result_data, d_broadcasted_shape, d_strides1, d_strides2, max_ndim, broadcasted_size);
+
+    cudaError_t error = cudaGetLastError();
+    if (error != cudaSuccess) {
+        printf("CUDA error: %s\n", cudaGetErrorString(error));
+        exit(-1);
+    }
+
+    cudaDeviceSynchronize();
+    cudaFree(d_broadcasted_shape);
+}
+
 __global__ void elementwise_mul_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int size) {
     
     int i = blockIdx.x * blockDim.x + threadIdx.x;
@@ -481,6 +542,89 @@ __host__ void log_tensor_cuda(Tensor* tensor, float* result_data) {
     cudaDeviceSynchronize();
 }
 
+__global__ void equal_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int size) {
+    
+    int i = blockIdx.x * blockDim.x + threadIdx.x;
+    if (i < size) {
+        result_data[i] = (data1[i] == data2[i]) ? 1.0f : 0.0f;
+    }
+}
+
+__host__ void equal_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data) {
+    
+    int number_of_blocks = (tensor1->size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
+    equal_tensor_cuda_kernel<<<number_of_blocks, THREADS_PER_BLOCK>>>(tensor1->data, tensor2->data, result_data, tensor1->size);
+
+    cudaError_t error = cudaGetLastError();
+    if (error != cudaSuccess) {
+        printf("CUDA error: %s\n", cudaGetErrorString(error));
+        exit(-1);
+    }
+
+    cudaDeviceSynchronize();
+}
+
+__global__ void equal_broadcasted_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int* broadcasted_shape, int* strides1, int*strides2, int max_ndim, int size) {
+    int i = blockIdx.x * blockDim.x + threadIdx.x;
+    if (i >= size) return;
+    
+    int index1 = 0, index2 = 0;
+    int linear_index = i;
+    for (int j = max_ndim - 1; j >= 0; j--) {
+        int pos = linear_index % broadcasted_shape[j];
+        linear_index /= broadcasted_shape[j];
+        if (strides1[j] != 0) index1 += pos * strides1[j];
+        if (strides2[j] != 0) index2 += pos * strides2[j];
+    }
+    result_data[i] = (data1[index1] == data2[index2]) ? 1.0f : 0.0f;
+}
+
+__host__ void equal_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size) {
+    int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
+
+    int* strides1 = (int*)malloc(max_ndim * sizeof(int));
+    int* strides2 = (int*)malloc(max_ndim * sizeof(int));
+    if (strides1 == NULL || strides2 == NULL) {
+        fprintf(stderr, "Memory allocation failed\n");
+        exit(1);
+    }
+
+    int stride1 = 1, stride2 = 1;
+    for (int i = max_ndim - 1; i >= 0; i--) {
+        int dim1 = i < tensor1->ndim ? tensor1->shape[tensor1->ndim - max_ndim + i] : 1;
+        int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - max_ndim + i] : 1;
+        strides1[i] = dim1 == broadcasted_shape[i] ? stride1 : 0;
+        strides2[i] = dim2 == broadcasted_shape[i] ? stride2 : 0;
+        stride1 *= (dim1 == broadcasted_shape[i]) ? dim1 : 1;
+        stride2 *= (dim2 == broadcasted_shape[i]) ? dim2 : 1;
+    }
+    
+    int* d_broadcasted_shape;
+    int* d_strides1;
+    int* d_strides2;
+
+    cudaMalloc((void**)&d_broadcasted_shape, max_ndim * sizeof(int));
+    cudaMemcpy(d_broadcasted_shape, broadcasted_shape, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
+
+    cudaMalloc((void**)&d_strides1, max_ndim * sizeof(int));
+    cudaMemcpy(d_strides1, strides1, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
+
+    cudaMalloc((void**)&d_strides2, max_ndim * sizeof(int));
+    cudaMemcpy(d_strides2, strides2, max_ndim * sizeof(int), cudaMemcpyHostToDevice);
+
+    int number_of_blocks = (broadcasted_size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
+    equal_broadcasted_tensor_cuda_kernel<<<number_of_blocks, THREADS_PER_BLOCK>>>(tensor1->data, tensor2->data, result_data, d_broadcasted_shape, d_strides1, d_strides2, max_ndim, broadcasted_size);
+
+    cudaError_t error = cudaGetLastError();
+    if (error != cudaSuccess) {
+        printf("CUDA error: %s\n", cudaGetErrorString(error));
+        exit(-1);
+    }
+
+    cudaDeviceSynchronize();
+    cudaFree(d_broadcasted_shape);
+}
+
 
 __global__ void ones_like_tensor_cuda_kernel(float* data, float* result_data, int size) {
     
diff --git a/norch/csrc/cuda.h b/norch/csrc/cuda.h
index 5d5e353..73eb831 100644
--- a/norch/csrc/cuda.h
+++ b/norch/csrc/cuda.h
@@ -8,7 +8,10 @@
     __host__ void add_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data);
 
     __global__ void add_broadcasted_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int* broadcasted_shape, int* strides1, int*strides2, int max_ndim, int size);
-    __host__ void add_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape);
+    __host__ void add_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size);
+
+    __global__ void sub_broadcasted_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int* broadcasted_shape, int* strides1, int*strides2, int max_ndim, int size);
+    __host__ void sub_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size);
 
     __global__ void sum_tensor_cuda_kernel(float* data, float* result_data);
     __host__ void sum_tensor_cuda(Tensor* tensor, float* result_data);
@@ -46,6 +49,12 @@
     __global__ void log_tensor_cuda_kernel(float* data, float* result_data, int size);
     __host__ void log_tensor_cuda(Tensor* tensor, float* result_data);
 
+    __global__ void equal_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int size);
+    __host__ void equal_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data);
+
+    __global__ void equal_broadcasted_tensor_cuda_kernel(float* data1, float* data2, float* result_data, int* broadcasted_shape, int* strides1, int*strides2, int max_ndim, int size);
+    __host__ void equal_broadcasted_tensor_cuda(Tensor* tensor1, Tensor* tensor2, float* result_data, int* broadcasted_shape, int broadcasted_size);
+
     __global__ void ones_like_tensor_cuda_kernel(float* data, float* result_data, int size);
     __host__ void ones_like_tensor_cuda(Tensor* tensor, float* result_data);
 
diff --git a/norch/csrc/tensor.cpp b/norch/csrc/tensor.cpp
index 104994e..889fff6 100644
--- a/norch/csrc/tensor.cpp
+++ b/norch/csrc/tensor.cpp
@@ -149,26 +149,30 @@ extern "C" {
             broadcasted_shape[max_ndim - 1 - i] = dim1 > dim2 ? dim1 : dim2;
         }
 
+        int broadcasted_size = 1;
+        for (int i = 0; i < max_ndim; i++) {
+            broadcasted_size *= broadcasted_shape[i];
+        }
+
         if (strcmp(tensor1->device, "cuda") == 0) {
             float* result_data;
-            cudaMalloc((void **)&result_data, tensor1->size * sizeof(float));
-            add_broadcasted_tensor_cuda(tensor1, tensor2, result_data, broadcasted_shape);
+            cudaMalloc((void **)&result_data, broadcasted_size * sizeof(float));
+            add_broadcasted_tensor_cuda(tensor1, tensor2, result_data, broadcasted_shape, broadcasted_size);
             return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
         } 
         else {
-            float* result_data = (float*)malloc(tensor1->size * sizeof(float));
+            float* result_data = (float*)malloc(broadcasted_size * sizeof(float));
             if (result_data == NULL) {
                 fprintf(stderr, "Memory allocation failed\n");
                 exit(1);
             }
 
-            add_broadcasted_tensor_cpu(tensor1, tensor2, result_data, broadcasted_shape);
+            add_broadcasted_tensor_cpu(tensor1, tensor2, result_data, broadcasted_shape, broadcasted_size);
             return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
         }
     }
 
-    Tensor* sum_tensor(Tensor* tensor, int axis) {
-
+    Tensor* sum_tensor(Tensor* tensor, int axis, bool keepdim) {
         char* device = (char*)malloc(strlen(tensor->device) + 1);
         if (device != NULL) {
             strcpy(device, tensor->device);
@@ -179,11 +183,8 @@ extern "C" {
         int ndim;
         int* shape;
         if (axis == -1) {
+            
             shape = (int*) malloc(sizeof(int));
-            if (shape == NULL) {
-                fprintf(stderr, "Memory allocation failed\n");
-                exit(1);
-            }
             shape[0] = 1;
             ndim = 1;
         } else {
@@ -195,21 +196,179 @@ extern "C" {
             }
             ndim = tensor->ndim - 1;
         }
+
+        int size = 1;
+        for (int i = 0; i < ndim; i++) {
+            size *= shape[i];
+        }
   
         if (strcmp(tensor->device, "cuda") == 0) {
 
             float* result_data;
-            cudaMalloc((void**)&result_data, tensor->size * sizeof(float));
+            cudaMalloc((void**)&result_data, size * sizeof(float));
+            cudaMemset(result_data, 0, size * sizeof(float));
             sum_tensor_cuda(tensor, result_data);
             return create_tensor(result_data, shape, ndim, device);
         } 
         else {
-            float* result_data = (float*)malloc(1 * sizeof(float));
+            float* result_data = (float*)calloc(size, sizeof(float));
             if (result_data == NULL) {
                 fprintf(stderr, "Memory allocation failed\n");
                 exit(1);
             }
-            sum_tensor_cpu(tensor, result_data, axis);
+
+            sum_tensor_cpu(tensor, result_data, size, shape, axis);
+
+            if (keepdim) {
+                if (axis == -1){
+                    ndim = tensor->ndim;
+                    shape = (int*) malloc((tensor->ndim) * sizeof(int));
+                    for (int i = 0; i < tensor->ndim; i++) {
+                        shape[i] = 1;
+                    }
+                } else {
+                    shape = (int*) malloc((tensor->ndim) * sizeof(int));
+                    for (int i = 0; i < tensor->ndim; i++) {
+                        shape[i] = tensor->shape[i];
+                    }
+                    shape[axis] = 1;
+                    ndim = tensor->ndim;
+                }
+                
+            }
+            return create_tensor(result_data, shape, ndim, device);
+        }     
+    }
+
+    Tensor* max_tensor(Tensor* tensor, int axis, bool keepdim) {
+        char* device = (char*)malloc(strlen(tensor->device) + 1);
+        if (device != NULL) {
+            strcpy(device, tensor->device);
+        } else {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(-1);
+        }
+        int ndim;
+        int* shape;
+        if (axis == -1) {     
+            shape = (int*) malloc(sizeof(int));
+            shape[0] = 1;
+            ndim = 1;
+        } else {
+            shape = (int*) malloc((tensor->ndim - 1) * sizeof(int));
+            for (int i = 0, j = 0; i < tensor->ndim; ++i) {
+                if (i != axis) {
+                    shape[j++] = tensor->shape[i];
+                }
+            }
+            ndim = tensor->ndim - 1;
+        }
+
+        int size = 1;
+        for (int i = 0; i < ndim; i++) {
+            size *= shape[i];
+        }
+  
+        if (strcmp(tensor->device, "cuda") == 0) {
+
+            float* result_data;
+            cudaMalloc((void**)&result_data, size * sizeof(float));
+            //cudaMemset(result_data, -INFINITY, size * sizeof(float));
+            //max_tensor_cuda(tensor, result_data);
+            return create_tensor(result_data, shape, ndim, device);
+        } 
+        else {
+            float* result_data = (float*)malloc(size * sizeof(float));
+            if (result_data == NULL) {
+                fprintf(stderr, "Memory allocation failed\n");
+                exit(1);
+            }
+            max_tensor_cpu(tensor, result_data, size, shape, axis);
+
+            if (keepdim) {
+                if (axis == -1){
+                    ndim = tensor->ndim;
+                    shape = (int*) malloc((tensor->ndim) * sizeof(int));
+                    for (int i = 0; i < tensor->ndim; i++) {
+                        shape[i] = 1;
+                    }
+                } else {
+                    shape = (int*) malloc((tensor->ndim) * sizeof(int));
+                    for (int i = 0; i < tensor->ndim; i++) {
+                        shape[i] = tensor->shape[i];
+                    }
+                    shape[axis] = 1;
+                    ndim = tensor->ndim;
+                }   
+            }
+
+            return create_tensor(result_data, shape, ndim, device);
+        }     
+    }
+
+    Tensor* min_tensor(Tensor* tensor, int axis, bool keepdim) {
+        char* device = (char*)malloc(strlen(tensor->device) + 1);
+        if (device != NULL) {
+            strcpy(device, tensor->device);
+        } else {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(-1);
+        }
+        int ndim;
+        int* shape;
+        if (axis == -1) {
+            
+            shape = (int*) malloc(sizeof(int));
+            shape[0] = 1;
+            ndim = 1;
+        } else {
+            shape = (int*) malloc((tensor->ndim - 1) * sizeof(int));
+            for (int i = 0, j = 0; i < tensor->ndim; ++i) {
+                if (i != axis) {
+                    shape[j++] = tensor->shape[i];
+                }
+            }
+            ndim = tensor->ndim - 1;
+        }
+        
+        int size = 1;
+        for (int i = 0; i < ndim; i++) {
+            size *= shape[i];
+        }
+  
+        if (strcmp(tensor->device, "cuda") == 0) {
+
+            float* result_data;
+            cudaMalloc((void**)&result_data, size * sizeof(float));
+            //cudaMemset(result_data, INFINITY, size * sizeof(float));
+            //min_tensor_cuda(tensor, result_data);
+            return create_tensor(result_data, shape, ndim, device);
+        } 
+        else {
+            float* result_data = (float*)malloc(size * sizeof(float));
+            if (result_data == NULL) {
+                fprintf(stderr, "Memory allocation failed\n");
+                exit(1);
+            }
+            min_tensor_cpu(tensor, result_data, size, shape, axis);
+
+            if (keepdim) {
+                if (axis == -1){
+                    ndim = tensor->ndim;
+                    shape = (int*) malloc((tensor->ndim) * sizeof(int));
+                    for (int i = 0; i < tensor->ndim; i++) {
+                        shape[i] = 1;
+                    }
+                } else {
+                    shape = (int*) malloc((tensor->ndim) * sizeof(int));
+                    for (int i = 0; i < tensor->ndim; i++) {
+                        shape[i] = tensor->shape[i];
+                    }
+                    shape[axis] = 1;
+                    ndim = tensor->ndim;
+                }   
+            }
+
             return create_tensor(result_data, shape, ndim, device);
         }     
     }
@@ -290,16 +449,27 @@ extern "C" {
             broadcasted_shape[max_ndim - 1 - i] = dim1 > dim2 ? dim1 : dim2;
         }
 
-        // Allocate memory for result tensor
-        float* result_data = (float*)malloc(tensor1->size * sizeof(float));
-        if (result_data == NULL) {
-            fprintf(stderr, "Memory allocation failed\n");
-            exit(1);
+        int broadcasted_size = 1;
+        for (int i = 0; i < max_ndim; i++) {
+            broadcasted_size *= broadcasted_shape[i];
         }
 
-        sub_broadcasted_tensor_cpu(tensor1, tensor2, result_data, broadcasted_shape);
+        if (strcmp(tensor1->device, "cuda") == 0) {
+            float* result_data;
+            cudaMalloc((void **)&result_data, broadcasted_size * sizeof(float));
+            sub_broadcasted_tensor_cuda(tensor1, tensor2, result_data, broadcasted_shape, broadcasted_size);
+            return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
+        } 
+        else {
+            float* result_data = (float*)malloc(broadcasted_size * sizeof(float));
+            if (result_data == NULL) {
+                fprintf(stderr, "Memory allocation failed\n");
+                exit(1);
+            }
 
-        return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
+            sub_broadcasted_tensor_cpu(tensor1, tensor2, result_data, broadcasted_shape, broadcasted_size);
+            return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
+        }
     }
 
     Tensor* elementwise_mul_tensor(Tensor* tensor1, Tensor* tensor2) {
@@ -469,7 +639,7 @@ extern "C" {
 
     Tensor* tensor_div_tensor(Tensor* tensor1, Tensor* tensor2) {
         if (tensor1->ndim != tensor2->ndim) {
-            fprintf(stderr, "Tensors must have the same number of dimensions %d and %d for element-wise multiplication\n", tensor1->ndim, tensor2->ndim);
+            fprintf(stderr, "Tensors must have the same number of dimensions %d and %d for element-wise division\n", tensor1->ndim, tensor2->ndim);
             exit(1);
         }
 
@@ -494,7 +664,7 @@ extern "C" {
 
         for (int i = 0; i < ndim; i++) {
             if (tensor1->shape[i] != tensor2->shape[i]) {
-                fprintf(stderr, "Tensors must have the same shape %d and %d at index %d for subtraction\n", tensor1->shape[i], tensor2->shape[i], i);
+                fprintf(stderr, "Tensors must have the same shape %d and %d at index %d for division\n", tensor1->shape[i], tensor2->shape[i], i);
                 exit(1);
             }
             shape[i] = tensor1->shape[i];
@@ -872,6 +1042,106 @@ extern "C" {
         }
     }
 
+    Tensor* equal_tensor(Tensor* tensor1, Tensor* tensor2) {
+        if (tensor1->ndim != tensor2->ndim) {
+            fprintf(stderr, "Tensors must have the same number of dimensions %d and %d for equal\n", tensor1->ndim, tensor2->ndim);
+            exit(1);
+        }
+
+        if (strcmp(tensor1->device, tensor2->device) != 0) {
+            fprintf(stderr, "Tensors must be on the same device: %s and %s\n", tensor1->device, tensor2->device);
+            exit(1);
+        }
+
+        char* device = (char*)malloc(strlen(tensor1->device) + 1);
+        if (device != NULL) {
+            strcpy(device, tensor1->device);
+        } else {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(-1);
+        }
+        int ndim = tensor1->ndim;
+        int* shape = (int*)malloc(ndim * sizeof(int));
+        if (shape == NULL) {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(1);
+        }
+
+        for (int i = 0; i < ndim; i++) {
+            if (tensor1->shape[i] != tensor2->shape[i]) {
+                fprintf(stderr, "Tensors must have the same shape %d and %d at index %d for equal\n", tensor1->shape[i], tensor2->shape[i], i);
+                exit(1);
+            }
+            shape[i] = tensor1->shape[i];
+        }        
+        
+        if (strcmp(tensor1->device, "cuda") == 0) {
+
+            float* result_data;
+            cudaMalloc((void **)&result_data, tensor1->size * sizeof(float));
+            equal_tensor_cuda(tensor1, tensor2, result_data);
+            return create_tensor(result_data, shape, ndim, device);
+        } 
+        else {
+            float* result_data = (float*)malloc(tensor1->size * sizeof(float));
+            if (result_data == NULL) {
+                fprintf(stderr, "Memory allocation failed\n");
+                exit(1);
+            }
+            equal_tensor_cpu(tensor1, tensor2, result_data);
+            return create_tensor(result_data, shape, ndim, device);
+        }     
+    }
+
+    Tensor* equal_broadcasted_tensor(Tensor* tensor1, Tensor* tensor2) {
+
+        if (strcmp(tensor1->device, tensor2->device) != 0) {
+            fprintf(stderr, "Tensors must be on the same device: %s and %s\n", tensor1->device, tensor2->device);
+            exit(1);
+        }
+
+        int max_ndim = tensor1->ndim > tensor2->ndim ? tensor1->ndim : tensor2->ndim;
+
+        // Determine the broadcasted shape
+        int* broadcasted_shape = (int*)malloc(max_ndim * sizeof(int));
+        if (broadcasted_shape == NULL) {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(1);
+        }
+        for (int i = 0; i < max_ndim; i++) {
+            int dim1 = i < tensor1->ndim ? tensor1->shape[tensor1->ndim - 1 - i] : 1;
+            int dim2 = i < tensor2->ndim ? tensor2->shape[tensor2->ndim - 1 - i] : 1;
+            if (dim1 != dim2 && dim1 != 1 && dim2 != 1) {
+                fprintf(stderr, "Shapes are not compatible for broadcasting\n");
+                exit(1);
+            }
+            broadcasted_shape[max_ndim - 1 - i] = dim1 > dim2 ? dim1 : dim2;
+        }
+
+        int broadcasted_size = 1;
+        for (int i = 0; i < max_ndim; i++) {
+            broadcasted_size *= broadcasted_shape[i];
+        }
+
+        if (strcmp(tensor1->device, "cuda") == 0) {
+            float* result_data;
+            cudaMalloc((void **)&result_data, broadcasted_size * sizeof(float));
+            equal_broadcasted_tensor_cuda(tensor1, tensor2, result_data, broadcasted_shape, broadcasted_size);
+            return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
+        } 
+        else {
+            float* result_data = (float*)malloc(broadcasted_size * sizeof(float));
+            if (result_data == NULL) {
+                fprintf(stderr, "Memory allocation failed\n");
+                exit(1);
+            }
+
+            equal_broadcasted_tensor_cpu(tensor1, tensor2, result_data, broadcasted_shape, broadcasted_size);
+            return create_tensor(result_data, broadcasted_shape, max_ndim, tensor1->device);
+        }
+    }
+
+
     Tensor* ones_like_tensor(Tensor* tensor) {
         char* device = (char*)malloc(strlen(tensor->device) + 1);
         if (device != NULL) {
diff --git a/norch/csrc/tensor.h b/norch/csrc/tensor.h
index 5ff62d0..5669d4f 100644
--- a/norch/csrc/tensor.h
+++ b/norch/csrc/tensor.h
@@ -16,7 +16,9 @@ extern "C" {
     Tensor* create_tensor(float* data, int* shape, int ndim, char* device);
     float get_item(Tensor* tensor, int* indices);
     Tensor* add_tensor(Tensor* tensor1, Tensor* tensor2);
-    Tensor* sum_tensor(Tensor* tensor, int axis);
+    Tensor* sum_tensor(Tensor* tensor, int axis, bool keepdims);
+    Tensor* max_tensor(Tensor* tensor, int axis, bool keepdim);
+    Tensor* min_tensor(Tensor* tensor, int axis, bool keepdim);
     Tensor* sub_tensor(Tensor* tensor1, Tensor* tensor2);
     Tensor* elementwise_mul_tensor(Tensor* tensor1, Tensor* tensor2);
     Tensor* scalar_mul_tensor(Tensor* tensor, float scalar);
@@ -28,6 +30,8 @@ extern "C" {
     Tensor* tensor_pow_scalar(Tensor* tensor, float exponent);
     Tensor* scalar_pow_tensor(float base, Tensor* tensor);
     Tensor* log_tensor(Tensor* tensor);
+    Tensor* equal_tensor(Tensor* tensor1, Tensor* tensor2);
+    Tensor* equal_broadcasted_tensor(Tensor* tensor1, Tensor* tensor2);
     void to_device(Tensor* tensor, char* device);
     Tensor* ones_like_tensor(Tensor* tensor);
     Tensor* zeros_like_tensor(Tensor* tensor);
diff --git a/norch/datasets/__init__.py b/norch/datasets/__init__.py
new file mode 100644
index 0000000..fe34b11
--- /dev/null
+++ b/norch/datasets/__init__.py
@@ -0,0 +1 @@
+from .mnist import *
\ No newline at end of file
diff --git a/norch/datasets/mnist.py b/norch/datasets/mnist.py
new file mode 100644
index 0000000..ff11edf
--- /dev/null
+++ b/norch/datasets/mnist.py
@@ -0,0 +1,90 @@
+import gzip
+import os
+import norch
+from norch.utils.data import Dataset
+import numpy as np
+
+class MNIST(Dataset):
+    """
+    Loads training, validation, and test partitions of the mnist dataset
+    (http://yann.lecun.com/exdb/mnist/). If the data is not already contained in data_dir, it will
+    try to download it.
+
+    This dataset contains 60000 training examples, and 10000 test examples of handwritten digits
+    in {0, ..., 9} and corresponding labels. Each handwritten image has an "original" dimension of
+    28x28x1, and is stored row-wise as a string of 784x1 bytes. Pixel values are in range 0 to 255
+    (inclusive).
+
+    Args:
+        data_dir: String. Relative or absolute path of the dataset.
+        devel_size: Integer. Size of the development (validation) dataset partition.
+
+    Returns:
+        X_train: float64 numpy array with shape [784, 60000-devel_size] with values in [0, 1].
+        Y_train: uint8 numpy array with shape [60000-devel_size]. Labels.
+        X_devel: float64 numpy array with shape [784, devel_size] with values in [0, 1].
+        Y_devel: uint8 numpy array with shape [devel_size]. Labels.
+        X_test: float64 numpy array with shape [784, 10000] with values in [0, 1].
+        Y_test: uint8 numpy array with shape [10000]. Labels.
+    """
+
+    urls = ['https://ossci-datasets.s3.amazonaws.com/mnist/train-images-idx3-ubyte.gz',
+            'https://ossci-datasets.s3.amazonaws.com/mnist/train-labels-idx1-ubyte.gz',
+            'https://ossci-datasets.s3.amazonaws.com/mnist/t10k-images-idx3-ubyte.gz',
+            'https://ossci-datasets.s3.amazonaws.com/mnist/t10k-labels-idx1-ubyte.gz',]
+    name = 'mnist-data-py'
+    dirname = 'mnist'
+
+    def __init__(self, path_data, path_label, transform=None, target_transform=None):
+        self.data = self._load_mnist(path_data, header_size=16).reshape((-1, 28, 28))
+        self.labels = self._load_mnist(path_label, header_size=8)
+        self.transform = transform
+        self.target_transform = target_transform
+
+    def _load_mnist(self, path, header_size):
+        with gzip.open(path, 'rb') as f:
+            data = np.frombuffer(f.read(), np.uint8, offset=header_size)
+        return np.asarray(data, dtype=np.uint8)
+
+
+    @classmethod
+    def splits(cls, root='.data', train_data='train-images-idx3-ubyte.gz', train_label='train-labels-idx1-ubyte.gz',
+               test_data='t10k-images-idx3-ubyte.gz', test_label='t10k-labels-idx1-ubyte.gz', **kwargs):
+        r"""
+        Loads training and test partitions of the [mnist dataset](https://www.cs.toronto.edu/~kriz/cifar.html). If
+        the data is not already contained in the ``root`` folder, it will download it.
+
+        Args:
+            root (str): relative or absolute path of the dataset.
+
+        Returns:
+            tuple(Dataset): training and testing datasets
+        """
+        path = os.path.join(root, cls.dirname, cls.name)
+        if not os.path.isdir(path):
+            path = cls.download(root)
+        train_data = os.path.join(path, train_data)
+        train_label = os.path.join(path, train_label)
+        test_data = os.path.join(path, test_data)
+        test_label = os.path.join(path, test_label)
+        return MNIST(train_data, train_label, **kwargs), MNIST(test_data, test_label, **kwargs)
+
+    def __getitem__(self, item):
+        data = self.data[item].tolist()
+        label = self.labels[item].tolist()
+
+        if self.transform is not None:
+            data = self.transform(data)
+        
+        if self.target_transform is not None:
+            label = self.target_transform(label)
+            
+        
+
+        return data, label
+
+    def __setitem__(self, key, value):
+        self.data[key], self.labels[key] = value
+
+    def __len__(self):
+        return len(self.data)
diff --git a/norch/libtensor.so b/norch/libtensor.so
index 06f1149..81254ad 100755
Binary files a/norch/libtensor.so and b/norch/libtensor.so differ
diff --git a/norch/nn/__init__.py b/norch/nn/__init__.py
index dadface..8f7997b 100644
--- a/norch/nn/__init__.py
+++ b/norch/nn/__init__.py
@@ -1,3 +1,4 @@
 from .modules import *
 from .activation import *
-from .loss import *
\ No newline at end of file
+from .loss import *
+from .functional import *
\ No newline at end of file
diff --git a/norch/nn/__pycache__/__init__.cpython-38.pyc b/norch/nn/__pycache__/__init__.cpython-38.pyc
index 1f19ab6..5b7fd1c 100644
Binary files a/norch/nn/__pycache__/__init__.cpython-38.pyc and b/norch/nn/__pycache__/__init__.cpython-38.pyc differ
diff --git a/norch/nn/__pycache__/activation.cpython-38.pyc b/norch/nn/__pycache__/activation.cpython-38.pyc
index f6d4dfe..0a2693b 100644
Binary files a/norch/nn/__pycache__/activation.cpython-38.pyc and b/norch/nn/__pycache__/activation.cpython-38.pyc differ
diff --git a/norch/nn/__pycache__/loss.cpython-38.pyc b/norch/nn/__pycache__/loss.cpython-38.pyc
index 170a29b..b8133b9 100644
Binary files a/norch/nn/__pycache__/loss.cpython-38.pyc and b/norch/nn/__pycache__/loss.cpython-38.pyc differ
diff --git a/norch/nn/activation.py b/norch/nn/activation.py
index 21cf891..a1361f6 100644
--- a/norch/nn/activation.py
+++ b/norch/nn/activation.py
@@ -1,4 +1,5 @@
 from .module import Module
+from . import functional as F
 import math
 
 class Activation(Module):
@@ -17,4 +18,13 @@ class Sigmoid(Activation):
         super().__init__()
 
     def forward(self, x):
-        return 1.0 / (1.0 + (math.e) ** (-x)) 
\ No newline at end of file
+        return F.sigmoid(x)
+    
+class Softmax(Activation):
+    def __init__(self, dim):
+        super(Softmax, self).__init__()
+
+        self.dim = dim
+
+    def forward(self, x):
+        return F.softmax(x, self.dim)
\ No newline at end of file
diff --git a/norch/nn/functional.py b/norch/nn/functional.py
new file mode 100644
index 0000000..9493591
--- /dev/null
+++ b/norch/nn/functional.py
@@ -0,0 +1,33 @@
+import math
+import norch
+import numpy as np
+
+def sigmoid(x):
+    return 1.0 / (1.0 + (math.e) ** (-x)) 
+
+def softmax(x, dim=None):
+    if dim is not None and dim < 0:
+        dim = x.ndim + dim
+            
+    x_max = x.max(axis=dim, keepdim=True)
+    exp_x = math.e ** (x - x_max)
+
+    if dim is not None:
+        sum_exp_x = exp_x.sum(axis=dim, keepdim=True) + exp_x.zeros_like()
+        return exp_x / sum_exp_x
+    else:
+        sum_exp_x = exp_x.sum()
+        return exp_x / sum_exp_x
+    
+def one_hot_encode(x, num_classes):
+    one_hot = [[0] * num_classes for _ in range(x.numel)]
+    
+    # Set the appropriate elements to 1
+    for i in range(x.numel):
+        target_idx = int(x.tensor.contents.data[i])
+        one_hot[i][target_idx] = 1
+
+    if x.numel < 2:
+        one_hot = one_hot[0]
+
+    return norch.Tensor(one_hot)
\ No newline at end of file
diff --git a/norch/nn/loss.py b/norch/nn/loss.py
index 789fab6..8d7b0f9 100644
--- a/norch/nn/loss.py
+++ b/norch/nn/loss.py
@@ -1,4 +1,6 @@
 from .module import Module
+from norch.autograd.functions import *
+import norch
 from abc import ABC
 
 class Loss(Module, ABC):
@@ -18,8 +20,66 @@ class MSELoss(Loss):
     def __init__(self):
         super().__init__()
 
-    def forward(self, predictions, labels):
-        assert labels.shape == predictions.shape, \
-            "Labels and predictions shape does not match: {} and {}".format(labels.shape, predictions.shape)
+    def forward(self, predictions, target):
+        assert target.shape == predictions.shape, \
+            "Labels and predictions shape does not match: {} and {}".format(target.shape, predictions.shape)
         
-        return ((predictions - labels) ** 2).sum() / predictions.numel
\ No newline at end of file
+        cost = ((predictions - target) ** 2).sum() / predictions.numel
+        return cost
+    
+
+class CrossEntropyLoss(Loss):
+    def __init__(self):
+        super().__init__()
+
+    def forward(self, input, target):
+        assert isinstance(input, norch.Tensor), \
+            "Cross entropy argument 'input' must be Tensor, not {}".format(type(input))
+        
+        assert isinstance(target, norch.Tensor), \
+            "Cross entropy argument 'target' must be Tensor, not {}".format(type(target))
+        
+        if input.ndim == 1:
+            if target.numel == 1:
+                num_classes = input.shape[0]
+                target = norch.one_hot_encode(target, num_classes)
+                
+                logits = norch.softmax(input, dim=0)
+                cost = -(logits.log() * target).sum()
+                
+            else:
+                # target -> class probabilities (one-hot encoded)
+                assert target.shape == input.shape, \
+                    "Input and target shape does not match: {} and {}".format(input.shape, target.shape)
+                logits = norch.softmax(input, dim=0)
+                cost = -(logits.log() * target).sum()
+
+
+        elif input.ndim == 2:
+            # batched 
+            if target.ndim == 1:
+                # target -> Ground truth class indices:
+                num_classes = input.shape[1]
+
+                target = norch.one_hot_encode(target, num_classes)
+                
+                batch_size = input.shape[0]
+                logits = norch.softmax(input, dim=1)
+                cost = -(logits.log() * target).sum() / batch_size
+
+            else:
+                # target -> class probabilities (one-hot encoded)
+                assert target.shape == input.shape, \
+                    "Input and target shape does not match: {} and {}".format(input.shape, target.shape)
+                
+                batch_size = input.shape[0]
+                logits = norch.softmax(input, dim=1)
+                cost = -(logits.log() * target).sum() / batch_size
+
+        if input.requires_grad:
+            cost.grad_fn = CrossEntropyLossBackward(input, target)
+
+        return cost
+            
+
+
diff --git a/norch/nn/modules/__pycache__/linear.cpython-38.pyc b/norch/nn/modules/__pycache__/linear.cpython-38.pyc
index d4a4c2c..583d8b6 100644
Binary files a/norch/nn/modules/__pycache__/linear.cpython-38.pyc and b/norch/nn/modules/__pycache__/linear.cpython-38.pyc differ
diff --git a/norch/nn/modules/linear.py b/norch/nn/modules/linear.py
index 2d61023..b90d845 100644
--- a/norch/nn/modules/linear.py
+++ b/norch/nn/modules/linear.py
@@ -2,15 +2,23 @@ from ..module import Module
 from ..parameter import Parameter
 
 class Linear(Module):
-    def __init__(self, input_dim, output_dim):
+    def __init__(self, input_dim, output_dim, bias=True):
         super().__init__()
         self.input_dim = input_dim
         self.output_dim = output_dim
         self.weight = Parameter(shape=[self.output_dim, self.input_dim])
-        self.bias = Parameter(shape=[self.output_dim, 1])
+
+        if bias:
+            self.bias = Parameter(shape=[self.output_dim, 1])
+        else:
+            self.bias = None
 
     def forward(self, x):
-        z = self.weight @ x + self.bias
+        if self.bias:
+            z = self.weight @ x + self.bias
+        else:
+            z = self.weight @ x
+
         return z
 
     def inner_repr(self):
diff --git a/norch/tensor.py b/norch/tensor.py
index 03b37ee..1121066 100644
--- a/norch/tensor.py
+++ b/norch/tensor.py
@@ -19,13 +19,18 @@ class Tensor:
     def __init__(self, data=None, device="cpu", requires_grad=False):
 
         if data != None:
+            if isinstance(data, (float, int)):
+                data = [data]
+
             data, shape = self.flatten(data)
-            self.data_ctype = (ctypes.c_float * len(data))(*data)
-            self.shape_ctype = (ctypes.c_int * len(shape))(*shape)
+            
+            self.shape = shape.copy()
+            
+            self.data_ctype = (ctypes.c_float * len(data))(*data.copy())
+            self.shape_ctype = (ctypes.c_int * len(shape))(*shape.copy())
             self.ndim_ctype = ctypes.c_int(len(shape))
             self.device_ctype = device.encode('utf-8')
 
-            self.shape = shape
             self.ndim = len(shape)
             self.device = device
 
@@ -109,6 +114,25 @@ class Tensor:
         return result_data
     
     def reshape(self, new_shape):
+        # Calculate the total number of elements in the tensor
+        total_elements = self.numel
+        
+        # Check for the presence of -1 in new_shape
+        if new_shape.count(-1) > 1:
+            raise ValueError("Only one dimension can be inferred (set to -1).")
+        
+        inferred_dim = None
+        known_dims_product = 1
+        for dim in new_shape:
+            if dim == -1:
+                inferred_dim = dim
+            else:
+                known_dims_product *= dim
+        
+        # Calculate the inferred dimension if -1 is present
+        if inferred_dim == -1:
+            inferred_dim_size = total_elements // known_dims_product
+            new_shape = [inferred_dim_size if dim == -1 else dim for dim in new_shape]
 
         new_shape_ctype = (ctypes.c_int * len(new_shape))(*new_shape)
         new_ndim_ctype = ctypes.c_int(len(new_shape))
@@ -129,7 +153,41 @@ class Tensor:
             result_data.grad_fn = ReshapeBackward(self)
 
         return result_data
-    
+
+    def unsqueeze(self, dim):
+        if dim < 0:
+            dim = self.ndim + dim + 1
+
+        # Ensure the dimension is valid
+        if dim > self.ndim:
+            raise ValueError("Dimension out of range (expected to be in range of [0, {0}], but got {1})".format(self.ndim, dim))
+        
+        # Create the new shape with an extra dimension of size 1
+        new_shape = self.shape[:dim] + [1] + self.shape[dim:]
+        
+        return self.reshape(new_shape)
+
+    def squeeze(self, dim=None):
+        if dim is not None:
+            if dim < 0:
+                dim = self.ndim + dim
+            
+            # Ensure the dimension is valid
+            if dim >= self.ndim or dim < 0:
+                raise ValueError("Dimension out of range (expected to be in range of [0, {0}), but got {1})".format(self.ndim, dim))
+            
+            # Only squeeze the specified dimension if its size is 1
+            if self.shape[dim] != 1:
+                raise ValueError("Dimension {0} does not have size 1 and cannot be squeezed".format(dim))
+            
+            # Create the new shape without the specified dimension
+            new_shape = self.shape[:dim] + self.shape[dim+1:]
+        else:
+            # Create the new shape by removing all dimensions of size 1
+            new_shape = [s for s in self.shape if s != 1]
+        
+        return self.reshape(new_shape)
+
     def to(self, device):
         self.device = device
         self.device_ctype = self.device.encode('utf-8')
@@ -224,6 +282,8 @@ class Tensor:
         if isinstance(other, (int, float)):
             other = other * self.ones_like()
 
+        broadcasted_shape_add = []
+
         # Function to determine if broadcasting is needed and get the broadcasted shape
         def broadcast_shape(shape1, shape2):
             if shape1 == shape2:
@@ -233,17 +293,25 @@ class Tensor:
             shape1 = [1] * (max_len - len(shape1)) + shape1
             shape2 = [1] * (max_len - len(shape2)) + shape2
 
-            broadcasted_shape = []
+            
             for dim1, dim2 in zip(shape1, shape2):
                 if dim1 != dim2 and dim1 != 1 and dim2 != 1:
                     raise ValueError("Shapes are not compatible for broadcasting")
-                broadcasted_shape.append(max(dim1, dim2))
-            return broadcasted_shape, True
+                broadcasted_shape_add.append(max(dim1, dim2))
+            return broadcasted_shape_add, True
 
-        broadcasted_shape, needs_broadcasting = broadcast_shape(self.shape, other.shape)
+        broadcasted_shape_add, needs_broadcasting = broadcast_shape(self.shape, other.shape)
 
         if needs_broadcasting:
             # Call add_broadcasted_tensor if broadcasting is needed
+            if other.ndim == self.ndim - 1:
+                other = other.reshape([1] + other.shape)
+            
+            elif self.ndim == other.ndim - 1:
+                self = self.reshape([1] + self.shape)
+                
+            
+                
             Tensor._C.add_broadcasted_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.POINTER(CTensor)]
             Tensor._C.add_broadcasted_tensor.restype = ctypes.POINTER(CTensor)
 
@@ -251,11 +319,13 @@ class Tensor:
 
             result_data = Tensor()
             result_data.tensor = result_tensor_ptr
-            result_data.shape = broadcasted_shape
-            result_data.ndim = len(broadcasted_shape)
+            result_data.shape = broadcasted_shape_add.copy()
+            result_data.ndim = len(broadcasted_shape_add)
 
             result_data.device = self.device
-            result_data.numel = self.numel  # Update this to calculate the correct number of elements if broadcasting
+            result_data.numel = 1
+            for s in result_data.shape:
+                result_data.numel *= s
 
             result_data.requires_grad = self.requires_grad or other.requires_grad
             if result_data.requires_grad:
@@ -312,6 +382,8 @@ class Tensor:
         if isinstance(other, (int, float)):
             other = other * self.ones_like()
 
+        broadcasted_shape_sub = []
+
         # Function to determine if broadcasting is needed and get the broadcasted shape
         def broadcast_shape(shape1, shape2):
             if shape1 == shape2:
@@ -321,16 +393,22 @@ class Tensor:
             shape1 = [1] * (max_len - len(shape1)) + shape1
             shape2 = [1] * (max_len - len(shape2)) + shape2
 
-            broadcasted_shape = []
+            
             for dim1, dim2 in zip(shape1, shape2):
                 if dim1 != dim2 and dim1 != 1 and dim2 != 1:
                     raise ValueError("Shapes are not compatible for broadcasting")
-                broadcasted_shape.append(max(dim1, dim2))
-            return broadcasted_shape, True
+                broadcasted_shape_sub.append(max(dim1, dim2))
+            return broadcasted_shape_sub, True
 
-        broadcasted_shape, needs_broadcasting = broadcast_shape(self.shape, other.shape)
+        broadcasted_shape_sub, needs_broadcasting = broadcast_shape(self.shape, other.shape)
 
         if needs_broadcasting:
+            if other.ndim == self.ndim - 1:
+                other = other.reshape([1] + other.shape)
+            
+            elif self.ndim == other.ndim - 1:
+                self = self.reshape([1] + self.shape)
+
             # Call add_broadcasted_tensor if broadcasting is needed
             Tensor._C.sub_broadcasted_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.POINTER(CTensor)]
             Tensor._C.sub_broadcasted_tensor.restype = ctypes.POINTER(CTensor)
@@ -339,8 +417,8 @@ class Tensor:
 
             result_data = Tensor()
             result_data.tensor = result_tensor_ptr
-            result_data.shape = broadcasted_shape
-            result_data.ndim = len(broadcasted_shape)
+            result_data.shape = broadcasted_shape_sub.copy()
+            result_data.ndim = len(broadcasted_shape_sub)
 
             result_data.device = self.device
             result_data.numel = self.numel  # Update this to calculate the correct number of elements if broadcasting
@@ -569,6 +647,9 @@ class Tensor:
                 result_data.grad_fn = DivisionBackward(self, other)
         
         elif isinstance(self, Tensor) and isinstance(other, Tensor):
+            if other.numel == 1:
+                return self.__truediv__(other.tensor.contents.data[0])
+            
             Tensor._C.tensor_div_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.POINTER(CTensor)]
             Tensor._C.tensor_div_tensor.restype = ctypes.POINTER(CTensor)
 
@@ -611,6 +692,80 @@ class Tensor:
         return result_data
 
     
+    def equal(self, other):
+
+        if isinstance(other, Tensor) and other.numel == 1:
+            # other is a single value tensor
+            other = self.zeros_like() + other
+            return self.equal(other)
+
+        if not isinstance(other, Tensor):
+            # other is a single value
+            if isinstance(other, (int, float)):
+                other = self.zeros_like() + other
+
+                return self.equal(other)
+            else:
+                return False
+        
+        broadcasted_shape_add = []
+
+        # Function to determine if broadcasting is needed and get the broadcasted shape
+        def broadcast_shape(shape1, shape2):
+            if shape1 == shape2:
+                return shape1, False
+            
+            max_len = max(len(shape1), len(shape2))
+            shape1 = [1] * (max_len - len(shape1)) + shape1
+            shape2 = [1] * (max_len - len(shape2)) + shape2
+
+            
+            for dim1, dim2 in zip(shape1, shape2):
+                if dim1 != dim2 and dim1 != 1 and dim2 != 1:
+                    raise ValueError("Shapes are not compatible for broadcasting")
+                broadcasted_shape_add.append(max(dim1, dim2))
+            return broadcasted_shape_add, True
+
+        broadcasted_shape_add, needs_broadcasting = broadcast_shape(self.shape, other.shape)
+        
+        if needs_broadcasting:
+            if other.ndim == self.ndim - 1:
+                other = other.reshape([1] + other.shape)
+            
+            elif self.ndim == other.ndim - 1:
+                self = self.reshape([1] + self.shape)
+
+            # Call equal_broadcasted_tensor if broadcasting is needed
+            Tensor._C.equal_broadcasted_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.POINTER(CTensor)]
+            Tensor._C.equal_broadcasted_tensor.restype = ctypes.POINTER(CTensor)
+
+            result_tensor_ptr = Tensor._C.equal_broadcasted_tensor(self.tensor, other.tensor)
+
+            result_data = Tensor()
+            result_data.tensor = result_tensor_ptr
+            result_data.shape = broadcasted_shape_add.copy()
+            result_data.ndim = len(broadcasted_shape_add)
+
+            result_data.device = self.device
+            result_data.numel = 1
+            for s in result_data.shape:
+                result_data.numel *= s
+
+        else:
+            Tensor._C.equal_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.POINTER(CTensor)]
+            Tensor._C.equal_tensor.restype = ctypes.POINTER(CTensor)
+
+            result_tensor_ptr = Tensor._C.equal_tensor(self.tensor, other.tensor)
+
+            result_data = Tensor()
+            result_data.tensor = result_tensor_ptr
+            result_data.shape = self.shape.copy()
+            result_data.ndim = self.ndim
+            result_data.device = self.device
+            result_data.numel = self.numel
+            
+        return result_data
+
     def log(self):
         Tensor._C.log_tensor.argtypes = [ctypes.POINTER(CTensor)]
         Tensor._C.log_tensor.restype = ctypes.POINTER(CTensor)
@@ -630,20 +785,34 @@ class Tensor:
 
         return result_data
     
-    def sum(self, axis=-1):
-        Tensor._C.sum_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.c_int]
+    def sum(self, axis=None, keepdim=False):
+        if axis is not None and axis < 0:
+            axis = self.ndim + axis
+            
+        if axis == None:
+            axis = -1
+
+        Tensor._C.sum_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.c_int, ctypes.c_bool]
         Tensor._C.sum_tensor.restype = ctypes.POINTER(CTensor)
 
-        result_tensor_ptr = Tensor._C.sum_tensor(self.tensor, axis)
+        result_tensor_ptr = Tensor._C.sum_tensor(self.tensor, axis, keepdim)
 
         result_data = Tensor()
         result_data.tensor = result_tensor_ptr
 
         if axis == -1:
-            result_data.shape = [1]
-            result_data.ndim = 1
+            if keepdim:
+                result_data.ndim = self.ndim
+                result_data.shape = [1] * self.ndim
+
+            else:
+                result_data.shape = [1]
+                result_data.ndim = 1
         else:
-            result_data.shape = self.shape[:axis] + self.shape[axis+1:]
+            if keepdim:
+                result_data.shape = self.shape[:axis] + [1] + self.shape[axis+1:]
+            else:
+                result_data.shape = self.shape[:axis] + self.shape[axis+1:]
             result_data.ndim = len(result_data.shape)
 
         result_data.device = self.device
@@ -653,7 +822,89 @@ class Tensor:
 
         result_data.requires_grad = self.requires_grad
         if result_data.requires_grad:
-            result_data.grad_fn = SumBackward(self)
+            result_data.grad_fn = SumBackward(self, axis, keepdim=keepdim)
+
+        return result_data
+    
+    def max(self, axis=None, keepdim=False):
+        if axis is not None and axis < 0:
+            axis = self.ndim + axis
+            
+        if axis == None:
+            axis = -1
+
+        Tensor._C.max_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.c_int, ctypes.c_bool]
+        Tensor._C.max_tensor.restype = ctypes.POINTER(CTensor)
+
+        result_tensor_ptr = Tensor._C.max_tensor(self.tensor, axis, keepdim)
+
+        result_data = Tensor()
+        result_data.tensor = result_tensor_ptr
+
+        if axis == -1:
+            if keepdim:
+                result_data.ndim = self.ndim
+                result_data.shape = [1] * self.ndim
+
+            else:
+                result_data.shape = [1]
+                result_data.ndim = 1
+        else:
+            if keepdim:
+                result_data.shape = self.shape[:axis] + [1] + self.shape[axis+1:]
+            else:
+                result_data.shape = self.shape[:axis] + self.shape[axis+1:]
+            result_data.ndim = len(result_data.shape)
+
+        result_data.device = self.device
+        result_data.numel = 1
+        for s in result_data.shape:
+            result_data.numel *= s
+
+        result_data.requires_grad = self.requires_grad
+        if result_data.requires_grad:
+            result_data.grad_fn = MaxBackward(self, axis, keepdim=keepdim)
+
+        return result_data
+    
+    def min(self, axis=None, keepdim=False):
+        if axis is not None and axis < 0:
+            axis = self.ndim + axis
+
+        if axis == None:
+            axis = -1
+        
+        Tensor._C.min_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.c_int, ctypes.c_bool]
+        Tensor._C.min_tensor.restype = ctypes.POINTER(CTensor)
+
+        result_tensor_ptr = Tensor._C.min_tensor(self.tensor, axis, keepdim)
+
+        result_data = Tensor()
+        result_data.tensor = result_tensor_ptr
+
+        if axis == -1:
+            if keepdim:
+                result_data.ndim = self.ndim
+                result_data.shape = [1] * self.ndim
+
+            else:
+                result_data.shape = [1]
+                result_data.ndim = 1
+        else:
+            if keepdim:
+                result_data.shape = self.shape[:axis] + [1] + self.shape[axis+1:]
+            else:
+                result_data.shape = self.shape[:axis] + self.shape[axis+1:]
+            result_data.ndim = len(result_data.shape)
+
+        result_data.device = self.device
+        result_data.numel = 1
+        for s in result_data.shape:
+            result_data.numel *= s
+
+        result_data.requires_grad = self.requires_grad
+        if result_data.requires_grad:
+            result_data.grad_fn = MinBackward(self, axis, keepdim=keepdim)
 
         return result_data
 
diff --git a/norch/utils/__init__.py b/norch/utils/__init__.py
index 90f60fd..c6adfd1 100644
--- a/norch/utils/__init__.py
+++ b/norch/utils/__init__.py
@@ -1 +1,2 @@
-from .utils import *
\ No newline at end of file
+from .utils import *
+from .data import *
\ No newline at end of file
diff --git a/norch/utils/__pycache__/utils.cpython-38.pyc b/norch/utils/__pycache__/utils.cpython-38.pyc
index 95d3b45..d78c80a 100644
Binary files a/norch/utils/__pycache__/utils.cpython-38.pyc and b/norch/utils/__pycache__/utils.cpython-38.pyc differ
diff --git a/norch/utils/data/__init__.py b/norch/utils/data/__init__.py
new file mode 100644
index 0000000..2397554
--- /dev/null
+++ b/norch/utils/data/__init__.py
@@ -0,0 +1,4 @@
+from .dataset import *
+from .example import *
+from .dataloader import *
+from .batch import *
\ No newline at end of file
diff --git a/norch/utils/data/batch.py b/norch/utils/data/batch.py
new file mode 100644
index 0000000..3c9b92c
--- /dev/null
+++ b/norch/utils/data/batch.py
@@ -0,0 +1,13 @@
+class Batch(object):
+    """
+    A ``Batch``depends on a ``Dataset`` and is made of ``batch_size`` ``Example``.
+    """
+    def __init__(self, example, batch_size):
+        self.example = example
+        self.batch_size = batch_size
+        
+    def __getitem__(self, item):
+        return self.example[item]
+
+    def __len__(self):
+        return self.batch_size
\ No newline at end of file
diff --git a/norch/utils/data/dataloader.py b/norch/utils/data/dataloader.py
new file mode 100644
index 0000000..01bd41e
--- /dev/null
+++ b/norch/utils/data/dataloader.py
@@ -0,0 +1,20 @@
+import numpy as np
+from .batch import Batch
+
+
+class Dataloader(object):
+
+    def __init__(self, dataset, batch_size=32):
+        self.dataset = dataset
+        self.batch_size = batch_size
+
+    def __iter__(self):
+        starts = np.arange(0, len(self.dataset), self.batch_size)
+
+        for start in starts:
+            end = start + self.batch_size
+            batch_size = min(end, len(self.dataset)) - start
+            yield Batch(self.dataset[start:end], batch_size)
+
+    def __len__(self):
+        return len(self.dataset) // self.batch_size
diff --git a/norch/utils/data/dataset.py b/norch/utils/data/dataset.py
new file mode 100644
index 0000000..d38c1ea
--- /dev/null
+++ b/norch/utils/data/dataset.py
@@ -0,0 +1,74 @@
+from abc import ABC, abstractmethod
+import os
+from norch.utils import extract_to_dir, download_from_url
+from .example import Example
+import norch
+
+
+class Dataset(ABC):
+    r"""
+    Abstract Dataset class. All dataset for machine learning purposes can inherits from this architecture,
+    for convenience.
+    """
+    urls = []
+    name = ''
+    dirname = ''
+
+    def __init__(self, examples, fields):
+        self.examples = examples
+        self.fields = fields
+
+    @classmethod
+    def splits(cls, train=None, test=None, valid=None, root='.'):
+        raise NotImplementedError
+
+    @classmethod
+    def download(cls, root):
+        r"""Download and unzip a web archive (.zip, .gz, or .tgz).
+
+        Args:
+            root (str): Folder to download data to.
+
+        Returns:
+            string: Path to extracted dataset.
+        """
+        path_dirname = os.path.join(root, cls.dirname)
+        path_name = os.path.join(path_dirname, cls.name)
+        if not os.path.isdir(path_dirname):
+            for url in cls.urls:
+                filename = os.path.basename(url)
+                zpath = os.path.join(path_dirname, filename)
+                if not os.path.isfile(zpath):
+                    if not os.path.exists(os.path.dirname(zpath)):
+                        os.makedirs(os.path.dirname(zpath))
+                    print(f'Download {filename} from {url} to {zpath}')
+                    download_from_url(url, zpath)
+                extract_to_dir(zpath, path_name)
+
+        return path_name
+
+    def __repr__(self):
+        name = self.__class__.__name__
+        string = f"Dataset {name}("
+        tab = "   "
+        for (key, value) in self.__dict__.items():
+            if key[0] != "_":
+                if isinstance(value, Example):
+                    fields = self.fields
+                    for (name, field) in fields:
+                        string += f"\n{tab}({name}): {field.__class__.__name__}" \
+                                  f"(transform={True if field.transform is not None else None}, dtype={field.dtype})"
+                elif isinstance(value, norch.Tensor):
+                    string += f"\n{tab}({key}): {value.__class__.__name__}(shape={value.shape}, dtype={value.dtype})"
+                else:
+                    string += f"\n{tab}({key}): {value.__class__.__name__}"
+        return f'{string}\n)'
+
+    def __getitem__(self, item):
+        return self.examples[item]
+
+    def __setitem__(self, key, value):
+        self.examples[key] = value
+
+    def __len__(self):
+        return len(self.examples)
diff --git a/norch/utils/data/example.py b/norch/utils/data/example.py
new file mode 100644
index 0000000..a4f4ebf
--- /dev/null
+++ b/norch/utils/data/example.py
@@ -0,0 +1,65 @@
+# File: example.py
+# Creation: Wednesday August 19th 2020
+# Author: Arthur Dujardin
+# Contact: arthur.dujardin@ensg.eu
+#          arthurd@ifi.uio.no
+# --------
+# Copyright (c) 2020 Arthur Dujardin
+
+
+class Field(object):
+    r"""
+    A ``Field`` defines the data to process from a raw dataset. It will convert the data into a tensor. The data can
+    be a string, integers, float etc. The data is meant to be preprocessed and the attribute ``transform`` handles
+    the way the user want to process the raw data.
+    """
+
+    def __init__(self, transform=None, dtype=None):
+        self.transform = transform
+        self.dtype = dtype
+
+    def process(self, value):
+        """Applies the transformation and changes the data's type if necessary.
+
+        Args:
+            value (Tensor):
+
+        Returns:
+
+        """
+        if self.transform is not None:
+            value = self.transform(value)
+        if self.dtype is not None:
+            value = value.astype(self.dtype)
+        return value
+
+
+class Example(object):
+    r"""
+    Store a single training / testing example, and store it as an attribute.
+    Highly inspired from PyTorch `Example <https://github.com/pytorch/text/blob/master/torchtext/data/example.py>`__.
+    
+    """
+    @classmethod
+    def fromlist(cls, values, fields):
+        """
+        Add an example from a list of data with respect to the fields.
+
+        Args:
+            values: raw data
+            fields (tuple(string, Field)): fields to preprocess the data on
+
+        Returns:
+            None
+        """
+        example = cls()
+        for (value, field) in zip(values, fields):
+            assert len(field) == 2, f"expected a field template similar to \
+                                    ('name_field', Field()) but got {format(field)}"
+            assert isinstance(field[1], Field), f"expected a field template similar to \
+                                                ('name_field', Field()) but got {format(field)}"
+            name = field[0]
+            value = field[1].process(value)
+            setattr(example, name, value)
+
+        return example
diff --git a/norch/utils/utils.py b/norch/utils/utils.py
index 1652e16..f8b9b30 100644
--- a/norch/utils/utils.py
+++ b/norch/utils/utils.py
@@ -1,4 +1,9 @@
 import random
+import requests
+import tarfile
+import zipfile
+import shutil
+import os
 
 def generate_random_list(shape):
     """
@@ -11,4 +16,105 @@ def generate_random_list(shape):
         if len(inner_shape) == 0:
             return [random.uniform(-1, 1) for _ in range(shape[0])]
         else:
-            return [generate_random_list(inner_shape) for _ in range(shape[0])]
\ No newline at end of file
+            return [generate_random_list(inner_shape) for _ in range(shape[0])]
+
+
+def download_from_url(url, save_path, chunk_size=128):
+    """Download a file from an URL.
+
+    Original answer from https://stackoverflow.com/questions/9419162/download-returned-zip-file-from-url
+
+    Args:
+        url (str): path to the URL.
+        save_path (str): path to the saving directory.
+        chunk_size (int): download chunk.
+
+    Returns:
+        None
+    """
+    response = requests.get(url, stream=True)
+    total = response.headers.get('content-length')
+    with open(save_path, 'wb') as f:
+        if total is None:
+            f.write(response.content)
+        else:
+            downloaded = 0
+            total = int(total)
+            for data in response.iter_content(chunk_size=max(int(total / 1000), 1024 * 1024)):
+                downloaded += len(data)
+                f.write(data)
+                progress_bar(downloaded, total, "Downloading...")
+
+
+def extract_to_dir(filename, dirpath='.'):
+    # Does not create folder twice with the same name
+    name, ext = os.path.splitext(filename)
+    # if os.path.basename(name) == os.path.basename(dirpath):
+    #     dirpath = '.'
+    # Extract
+    print(dirpath)
+    print("Extracting...", end="")
+    if tarfile.is_tarfile(filename):
+        tarfile.open(filename, 'r').extractall(dirpath)
+    elif zipfile.is_zipfile(filename):
+        zipfile.ZipFile(filename, 'r').extractall(dirpath)
+    elif ext == '.gz':
+        if not os.path.exists(dirpath):
+            os.mkdir(dirpath)
+        shutil.move(filename, os.path.join(dirpath, os.path.basename(filename)))
+        print(f" | NOTE: gzip files are not extracted, and moved to {dirpath}", end="")
+    # Return the path where the file was extracted
+    print(" | Done !")
+    return os.path.abspath(dirpath)
+
+def progress_bar(current_index, max_index, prefix=None, suffix=None, start_time=None):
+    """Display a progress bar and duration.
+
+    Args:
+        current_index (int): current state index (or epoch number).
+        max_index (int): maximal numbers of state.
+        prefix (str, optional): prefix of the progress bar. The default is None.
+        suffix (str, optional): suffix of the progress bar. The default is None.
+        start_time (float, optional): starting time of the progress bar. If not None, it will display the time
+            spent from the beginning to the current state. The default is None.
+
+    Returns:
+        None. Display the progress bar in the console.
+    """
+    # Add a prefix to the progress bar
+    prefix = "" if prefix is None else str(prefix) + " "
+
+    # Get the percentage
+    percentage = current_index * 100 // max_index
+    loading = "[" + "=" * (percentage // 2) + " " * (50 - percentage // 2) + "]"
+    progress_display = "\r{0}{1:3d}% | {2}".format(prefix, percentage, loading)
+
+    # Add a suffix to the progress bar
+    progress_display += "" if suffix is None else " | " + str(suffix)
+
+    # Add a timer
+    if start_time is not None:
+        time_min, time_sec = get_time(start_time, time.time())
+        time_display = " | Time: {0}m {1}s".format(time_min, time_sec)
+        progress_display += time_display
+
+    # Print the progress bar
+    # TODO: return a string instead
+    print(progress_display, end="{}".format("" if current_index < max_index else " | Done !\n"))
+
+def get_time(start_time, end_time):
+    """Get ellapsed time in minutes and seconds.
+
+    Args:
+        start_time (float): strarting time
+        end_time (float): ending time
+
+    Returns:
+        elapsed_mins (float): elapsed time in minutes
+        elapsed_secs (float): elapsed time in seconds.
+    """
+    elapsed_time = end_time - start_time
+    elapsed_mins = int(elapsed_time / 60)
+    elapsed_secs = int(elapsed_time - (elapsed_mins * 60))
+
+    return elapsed_mins, elapsed_secs
diff --git a/test.py b/test.py
new file mode 100644
index 0000000..7ad7e2d
--- /dev/null
+++ b/test.py
@@ -0,0 +1,75 @@
+import norch
+from norch.utils.data.dataloader import Dataloader
+import norch
+import norch.nn as nn
+import norch.optim as optim
+import random
+random.seed(1)
+
+
+to_tensor = lambda x: norch.Tensor(x)
+reshape = lambda x: x.reshape([-1, 784])
+transform = lambda x: reshape(to_tensor(x))
+target_transform = lambda x: to_tensor(x)
+
+train_data, test_data = norch.datasets.MNIST.splits(transform=transform, target_transform=target_transform)
+sample, _ = train_data[0]
+
+BATCH_SIZE = 100
+
+train_loader = Dataloader(train_data, batch_size = BATCH_SIZE)
+
+class MyModel(nn.Module):
+    def __init__(self):
+        super(MyModel, self).__init__()
+        self.fc1 = nn.Linear(784, 5)
+        self.sigmoid = nn.Sigmoid()
+        self.fc2 = nn.Linear(5, 10)
+
+    def forward(self, x):
+        out = self.fc1(x)
+        out = self.sigmoid(out)
+        out = self.fc2(out)
+        
+        return out
+
+device = "cpu"
+epochs = 10
+
+model = MyModel().to(device)
+criterion = nn.CrossEntropyLoss()
+optimizer = optim.SGD(model.parameters(), lr=0.001)
+loss_list = []
+
+for epoch in range(epochs):    
+    for idx, batch in enumerate(train_loader):
+        x, target = batch
+
+        x = x.unsqueeze(-1)
+        target = target
+
+        x = x.to(device)
+        target = target.to(device).unsqueeze(-1)
+
+        outputs = model(x)
+        print(outputs.shape, target.shape, x.shape)
+        loss = criterion(outputs, target)
+        
+        optimizer.zero_grad()
+        loss.backward()
+        print('loss_antes', loss)
+
+        print('f1 antes', model.fc1.bias)
+        print('f1 grad_antes', model.fc1.bias.grad)
+        print('f2 antes', model.fc2.bias)
+        print('f2 grad_antes', model.fc2.bias.grad)
+
+        optimizer.step()
+        print('\n')
+
+        print('f1 depois', model.fc1.bias)
+        print('f2 depois', model.fc2.bias)
+        print('\n\n')
+
+    print(f'Epoch [{epoch + 1}/{epochs}], Loss: {loss[0]:.4f}')
+    loss_list.append(loss[0])
\ No newline at end of file
diff --git a/tests/test_autograd.py b/tests/test_autograd.py
index 42431a8..4739e7c 100644
--- a/tests/test_autograd.py
+++ b/tests/test_autograd.py
@@ -34,7 +34,204 @@ class TestTensorAutograd(unittest.TestCase):
         self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
         self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
 
-    def test_broadcasting_addition_autograd(self):
+    def test_sum_axis(self):
+        """
+        Test autograd from sum specifying axis
+        """
+        norch_tensor1 = norch.Tensor([[[1, 2.5], [3, -4]], [[5, 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_tensor2 = norch.Tensor([[[1, 1.], [1, 1.9]], [[1, 1], [1, 1]]], requires_grad=True).to(self.device)
+        norch_result = (norch_tensor1 + norch_tensor2).sum(axis=0).sum(axis=0).sum()
+
+        norch_result.backward()
+        norch_tensor1_grad = utils.to_torch(norch_tensor1.grad).to(self.device)
+        norch_tensor2_grad = utils.to_torch(norch_tensor2.grad).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[1, 2.5], [3, -4]], [[5, 6], [7, 8]]], requires_grad=True).to(self.device)
+        torch_tensor2 = torch.tensor([[[1, 1.], [1, 1.9]], [[1, 1], [1, 1]]], requires_grad=True).to(self.device)
+        
+        torch_result = (torch_tensor1 + torch_tensor2).sum(axis=0).sum(axis=0).sum()
+        torch_result.backward()
+        torch_tensor1_grad = torch_tensor1.grad
+        torch_tensor2_grad = torch_tensor2.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
+        self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
+
+        norch_tensor1 = norch.Tensor([[[1, 2.5], [3, -4]], [[5, 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_tensor2 = norch.Tensor([[[1, 1.], [1, 1.9]], [[1, 1], [1, 1]]], requires_grad=True).to(self.device)
+        norch_result = (norch_tensor1 + norch_tensor2).sum(axis=1).sum()
+
+        norch_result.backward()
+        norch_tensor1_grad = utils.to_torch(norch_tensor1.grad).to(self.device)
+        norch_tensor2_grad = utils.to_torch(norch_tensor2.grad).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[1, 2.5], [3, -4]], [[5, 6], [7, 8]]], requires_grad=True).to(self.device)
+        torch_tensor2 = torch.tensor([[[1, 1.], [1, 1.9]], [[1, 1], [1, 1]]], requires_grad=True).to(self.device)
+        
+        torch_result = (torch_tensor1 + torch_tensor2).sum(axis=1).sum()
+        torch_result.backward()
+        torch_tensor1_grad = torch_tensor1.grad
+        torch_tensor2_grad = torch_tensor2.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
+        self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
+    
+    
+    def test_max(self):
+        """
+        Test autograd from max
+        """
+        norch_tensor = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_result = norch_tensor.max()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        
+        torch_result = torch_tensor.max()
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+    
+    def test_max_axis(self):
+        """
+        Test autograd from max specifying axis
+        """
+        norch_tensor = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_max_axis = norch_tensor.max(axis=1)
+        norch_result = norch_max_axis.sum()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+
+        torch_max_axis, _ = torch_tensor.max(axis=1)
+        torch_result = torch_max_axis.sum()
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+        norch_tensor = norch.Tensor([[[10, 1], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+        norch_max_axis = norch_tensor.max(axis=2)
+        norch_result = norch_max_axis.sum()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 1], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+
+        torch_max_axis, _ = torch_tensor.max(axis=2)
+        torch_result = torch_max_axis.sum()
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+
+    ## evaluate case with some repeated values and axis 2
+
+    #def test_max_axis(self):
+    #    """
+    #    Test autograd from max specifying axis
+    #    """
+    #
+    #    norch_tensor = norch.Tensor([[[10, 10], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+    #    norch_max_axis = norch_tensor.max(axis=2)
+    #    norch_result = norch_max_axis.sum()
+    #
+    #    norch_result.backward()
+    #    norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+    #
+    #    torch_tensor = torch.tensor([[[10, 10], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+    #
+    #    torch_max_axis, _ = torch_tensor.max(axis=2)
+    #    torch_result = torch_max_axis.sum()
+    #    torch_result.backward()
+    #    torch_tensor_grad = torch_tensor.grad
+    #    self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+
+    #def test_min_axis(self):
+    #    """
+    #    Test autograd from max specifying axis
+    #    """
+    #
+    #    norch_tensor = norch.Tensor([[[10, 10], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+    #    norch_min_axis = norch_tensor.min(axis=2)
+    #    norch_result = norch_min_axis.sum()
+    #
+    #    norch_result.backward()
+    #    norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+    #
+    #    torch_tensor = torch.tensor([[[10, 10], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+    #
+    #    torch_min_axis, _ = torch_tensor.min(axis=2)
+    #    torch_result = torch_min_axis.sum()
+    #    torch_result.backward()
+    #    torch_tensor_grad = torch_tensor.grad
+    #    self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+
+    
+    def test_min(self):
+        """
+        Test autograd from min
+        """
+        norch_tensor = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_result = norch_tensor.min()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        
+        torch_result = torch_tensor.min()
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+    def test_min_axis(self):
+        """
+        Test autograd from min specifying axis
+        """
+        norch_tensor = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_min = norch_tensor.min(axis=1)
+        norch_result = norch_min.sum()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        
+        torch_min, _ = torch_tensor.min(axis=1)
+        torch_result = torch_min.sum()
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+        norch_tensor = norch.Tensor([[[10, 1], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+        norch_min = norch_tensor.min(axis=2)
+        norch_result = norch_min.sum()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 1], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+        
+        torch_min, _ = torch_tensor.min(axis=2)
+        torch_result = torch_min.sum()
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+    
+
+    def test_broadcasted_addition_autograd(self):
         """
         Test autograd for broadcasting addition: tensor1 + tensor2
         """
@@ -54,6 +251,26 @@ class TestTensorAutograd(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
         self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
+
+        ## reversed order broadcasting
+        norch_tensor1 = norch.Tensor([[[1., 2, 3], [4, 5, 6]]], requires_grad=True).to(self.device)  # Shape (1, 2, 3)
+        norch_tensor2 = norch.Tensor([1.5, -1, 0], requires_grad=True).to(self.device)  # Shape (3)
+
+        norch_result = (norch_tensor2 + norch_tensor1).sum()
+        norch_result.backward()
+        norch_tensor1_grad = utils.to_torch(norch_tensor1.grad).to(self.device)
+        norch_tensor2_grad = utils.to_torch(norch_tensor2.grad).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[1., 2, 3], [4, 5, 6]]], requires_grad=True).to(self.device)  # Shape (1, 2, 3)
+        torch_tensor2 = torch.tensor([1.5, -1, 0], requires_grad=True).to(self.device)  # Shape (3)
+
+        torch_result = (torch_tensor2 + torch_tensor1).sum()
+        torch_result.backward()
+        torch_tensor1_grad = torch_tensor1.grad
+        torch_tensor2_grad = torch_tensor2.grad
+
+        self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
+        self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
     
     def test_subtraction(self):
         """
@@ -76,7 +293,7 @@ class TestTensorAutograd(unittest.TestCase):
         self.assertTrue(utils.compare_torch(norch_tensor1_grad_sub, torch_tensor1_grad_sub))
         self.assertTrue(utils.compare_torch(norch_tensor2_grad_sub, torch_tensor2_grad_sub))
         
-    def test_broadcasting_subtraction_autograd(self):
+    def test_broadcasted_subtraction_autograd(self):
         """
         Test autograd for broadcasting subtraction: tensor1 - tensor2
         """
@@ -96,6 +313,26 @@ class TestTensorAutograd(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
         self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
+
+        # reversed order broadcasting
+        norch_tensor1 = norch.Tensor([[[1., 2, 3], [4, 5, 6]]], requires_grad=True).to(self.device)  # Shape (1, 2, 3)
+        norch_tensor2 = norch.Tensor([1.5, -1, 0], requires_grad=True).to(self.device)  # Shape (3)
+
+        norch_result = (norch_tensor2 - norch_tensor1).sum()
+        norch_result.backward()
+        norch_tensor1_grad = utils.to_torch(norch_tensor1.grad).to(self.device)
+        norch_tensor2_grad = utils.to_torch(norch_tensor2.grad).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[1., 2, 3], [4, 5, 6]]], requires_grad=True).to(self.device)  # Shape (1, 2, 3)
+        torch_tensor2 = torch.tensor([1.5, -1, 0], requires_grad=True).to(self.device)  # Shape (3)
+        
+        torch_result = (torch_tensor2 - torch_tensor1).sum()
+        torch_result.backward()
+        torch_tensor1_grad = torch_tensor1.grad
+        torch_tensor2_grad = torch_tensor2.grad
+
+        self.assertTrue(utils.compare_torch(norch_tensor1_grad, torch_tensor1_grad))
+        self.assertTrue(utils.compare_torch(norch_tensor2_grad, torch_tensor2_grad))
     
 
     def test_division(self):
@@ -211,7 +448,71 @@ class TestTensorAutograd(unittest.TestCase):
         self.assertTrue(utils.compare_torch(norch_tensor1_grad_matmul, torch_tensor1_grad_matmul))
         self.assertTrue(utils.compare_torch(norch_tensor2_grad_matmul, torch_tensor2_grad_matmul))
     
-    
+    def test_batched_matmul(self):
+        """
+        Test autograd from batched matrix multiplication: BxMxP = BxNxM @ BxMxP
+        """
+        B = 3  # Batch size
+
+        norch_tensor1_matmul = norch.Tensor([[[1., 2], [3, -4], [5, 6], [7, 8]] for _ in range(B)], requires_grad=True).to(self.device)
+        norch_tensor2_matmul = norch.Tensor([[[2., 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)], requires_grad=True).to(self.device)
+
+        norch_result_matmul = norch_tensor1_matmul @ norch_tensor2_matmul
+        norch_result_matmul_sum = norch_result_matmul.sum()  # Sum over all elements
+        norch_result_matmul_sum.backward()
+
+        # Convert gradients to torch tensors
+        norch_tensor1_grad_matmul = utils.to_torch(norch_tensor1_matmul.grad).to(self.device)
+        norch_tensor2_grad_matmul = utils.to_torch(norch_tensor2_matmul.grad).to(self.device)
+
+        # Repeat the same process with torch tensors
+        torch_tensor1_matmul = torch.tensor([[[1., 2], [3, -4], [5, 6], [7, 8]] for _ in range(B)], requires_grad=True).to(self.device)
+        torch_tensor2_matmul = torch.tensor([[[2., 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)], requires_grad=True).to(self.device)
+        torch_result_matmul = torch.matmul(torch_tensor1_matmul, torch_tensor2_matmul)
+        torch_result_matmul_sum = torch_result_matmul.sum()
+        torch_result_matmul_sum.backward()
+        
+        # Extract gradients from torch tensors
+        torch_tensor1_grad_matmul = torch_tensor1_matmul.grad
+        torch_tensor2_grad_matmul = torch_tensor2_matmul.grad
+
+        # Assertions to compare the gradients
+        self.assertTrue(utils.compare_torch(norch_tensor1_grad_matmul, torch_tensor1_grad_matmul))
+        self.assertTrue(utils.compare_torch(norch_tensor2_grad_matmul, torch_tensor2_grad_matmul))
+
+    def test_broadcasted_batched_matmul(self):
+        """
+        Test autograd from broadcasted batched matrix multiplication: BxMxP = NxM @ BxMxP
+        """
+        B = 3  # Batch size
+
+        norch_tensor1_matmul = norch.Tensor([[1., 2], [3, -4], [5, 6], [7, 8]], requires_grad=True).to(self.device)
+        norch_tensor2_matmul = norch.Tensor([[[2., 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)], requires_grad=True).to(self.device)
+
+        norch_result_matmul = norch_tensor1_matmul @ norch_tensor2_matmul
+        norch_result_matmul_sum = norch_result_matmul.sum()  # Sum over all elements
+        norch_result_matmul_sum.backward()
+
+        # Convert gradients to torch tensors
+        norch_tensor1_grad_matmul = utils.to_torch(norch_tensor1_matmul.grad).to(self.device)
+        norch_tensor2_grad_matmul = utils.to_torch(norch_tensor2_matmul.grad).to(self.device)
+
+        # Repeat the same process with torch tensors
+        torch_tensor1_matmul = torch.tensor([[1., 2], [3, -4], [5, 6], [7, 8]], requires_grad=True).to(self.device)
+        torch_tensor2_matmul = torch.tensor([[[2., 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)], requires_grad=True).to(self.device)
+        torch_result_matmul = torch.matmul(torch_tensor1_matmul, torch_tensor2_matmul)
+        torch_result_matmul_sum = torch_result_matmul.sum()
+        torch_result_matmul_sum.backward()
+        
+        # Extract gradients from torch tensors
+        torch_tensor1_grad_matmul = torch_tensor1_matmul.grad
+        torch_tensor2_grad_matmul = torch_tensor2_matmul.grad
+
+        # Assertions to compare the gradients
+        self.assertTrue(utils.compare_torch(norch_tensor1_grad_matmul, torch_tensor1_grad_matmul))
+        self.assertTrue(utils.compare_torch(norch_tensor2_grad_matmul, torch_tensor2_grad_matmul))
+
+
     def test_elementwise_mul_scalar(self):
         """
         Test autograd from elementwise multiplication with scalar: scalar * tensor
@@ -282,7 +583,169 @@ class TestTensorAutograd(unittest.TestCase):
         torch_expected_cos_tensor_grad = torch_cos_tensor.grad
 
         self.assertTrue(utils.compare_torch(torch_result_cos_tensor_grad, torch_expected_cos_tensor_grad))
-    
+
+    def test_sigmoid(self):
+        """
+        Test autograd from sigmoid
+        """
+        norch_tensor = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        norch_sigmoid = norch.sigmoid(norch_tensor)
+        norch_result = norch_sigmoid.sum()
+
+        norch_result.backward()
+        norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+        torch_tensor = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+        torch_sigmoid = torch.sigmoid(torch_tensor)
+        torch_result = torch_sigmoid.sum()
+
+        torch_result.backward()
+        torch_tensor_grad = torch_tensor.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+    def test_mse_loss_autograd(self):
+        """
+        Test the MSELoss with autograd functionality
+        """
+        loss_fn_norch = norch.nn.MSELoss()
+        loss_fn_torch = torch.nn.MSELoss()
+
+        predictions_norch = norch.Tensor([1.1, 2, 3, 4], requires_grad=True).to(self.device)
+        labels_norch = norch.Tensor([4, 3, 2.1, 1]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_norch.backward()  # Backpropagate the loss
+        grad_norch = predictions_norch.grad
+
+        predictions_torch = torch.tensor([1.1, 2, 3, 4], requires_grad=True).to(self.device)
+        labels_torch = torch.tensor([4, 3, 2.1, 1]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        loss_torch_expected.backward()  # Backpropagate the loss
+        grad_torch_expected = predictions_torch.grad
+
+        # Convert norch gradient to torch tensor for comparison
+        grad_norch_torch = utils.to_torch(grad_norch).to(self.device)
+
+        self.assertTrue(utils.compare_torch(grad_norch_torch, grad_torch_expected))
+
+    def test_cross_entropy_loss_autograd(self):
+        """
+        Test the CrossEntropyLoss with autograd functionality
+        """
+        loss_fn_norch = norch.nn.CrossEntropyLoss()
+        loss_fn_torch = torch.nn.CrossEntropyLoss()
+
+        # Test case 1: Single class, single sample
+        predictions_norch = norch.Tensor([2.0, 1.0, 0.1], requires_grad=True).to(self.device)
+        labels_norch = norch.Tensor([0]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+
+        loss_norch.backward()  # Backpropagate the loss
+        grad_norch = predictions_norch.grad
+
+        predictions_torch = torch.tensor([2.0, 1.0, 0.1], requires_grad=True).to(self.device)
+        labels_torch = torch.tensor(0).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        loss_torch_expected.backward()  # Backpropagate the loss
+        grad_torch_expected = predictions_torch.grad
+
+        # Convert norch gradient to torch tensor for comparison
+        grad_norch_torch = utils.to_torch(grad_norch).to(self.device)
+
+        self.assertTrue(utils.compare_torch(grad_norch_torch, grad_torch_expected))
+
+        # Test case 2: Multiple classes, multiple samples
+        predictions_norch = norch.Tensor([[0.5, 1.5, 2.5], [1.0, 2.0, 3.0]], requires_grad=True).to(self.device)
+        labels_norch = norch.Tensor([2, 1]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_norch.backward()  # Backpropagate the loss
+        grad_norch = predictions_norch.grad
+
+        predictions_torch = torch.tensor([[0.5, 1.5, 2.5], [1.0, 2.0, 3.0]], requires_grad=True).to(self.device)
+        labels_torch = torch.tensor([2, 1]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        loss_torch_expected.backward()  # Backpropagate the loss
+        grad_torch_expected = predictions_torch.grad
+
+        # Convert norch gradient to torch tensor for comparison
+        grad_norch_torch = utils.to_torch(grad_norch).to(self.device)
+
+        self.assertTrue(utils.compare_torch(grad_norch_torch, grad_torch_expected))
+
+        # Test case 3: Edge case - all predictions are zero
+        predictions_norch = norch.Tensor([[0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], requires_grad=True).to(self.device)
+        labels_norch = norch.Tensor([1, 2]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_norch.backward()  # Backpropagate the loss
+        grad_norch = predictions_norch.grad
+
+        predictions_torch = torch.tensor([[0.0, 0.0, 0.0], [0.0, 0.0, 0.0]], requires_grad=True).to(self.device)
+        labels_torch = torch.tensor([1, 2]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        loss_torch_expected.backward()  # Backpropagate the loss
+        grad_torch_expected = predictions_torch.grad
+
+        # Convert norch gradient to torch tensor for comparison
+        grad_norch_torch = utils.to_torch(grad_norch).to(self.device)
+
+        self.assertTrue(utils.compare_torch(grad_norch_torch, grad_torch_expected))
+
+        # Test case 4: Batched class probabilities instead of class index
+        predictions_norch = norch.Tensor([[0.5, 0.2, 0.1], [0.1, 0.5, 0.7]], requires_grad=True).to(self.device)
+        labels_norch = norch.Tensor([[1., 0, 0], [0, 1, 0]]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_norch.backward()  # Backpropagate the loss
+        grad_norch = predictions_norch.grad
+
+        predictions_torch = torch.tensor([[0.5, 0.2, 0.1], [0.1, 0.5, 0.7]], requires_grad=True).to(self.device)
+        labels_torch = torch.tensor([[1., 0, 0], [0, 1, 0]]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        loss_torch_expected.backward()  # Backpropagate the loss
+        grad_torch_expected = predictions_torch.grad
+
+        # Convert norch gradient to torch tensor for comparison
+        grad_norch_torch = utils.to_torch(grad_norch).to(self.device)
+
+        self.assertTrue(utils.compare_torch(grad_norch_torch, grad_torch_expected))
+
+
+    # implement grad pure softmax --> 0
+    # def test_softmax(self):
+    #     """
+    #     Test autograd from softmax
+    #     """
+    #     norch_tensor = norch.Tensor([[[-5, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+    #     norch_softmax = norch.softmax(norch_tensor, dim=1)
+    #     norch_result = norch_softmax.sum()
+
+    #     norch_result.backward()
+    #     norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+    #     torch_tensor = torch.tensor([[[-5, 10], [-4, -4]], [[5., 6], [7, 8]]], requires_grad=True).to(self.device)
+    #     torch_softmax = torch.softmax(torch_tensor, dim=1)
+    #     torch_result = torch_softmax.sum()
+
+    #     torch_result.backward()
+    #     torch_tensor_grad = torch_tensor.grad
+        
+    #     self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
+    #     norch_tensor = norch.Tensor([[[10, 1], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+    #     norch_softmax = norch.softmax(norch_tensor, dim=2)
+    #     norch_result = norch_softmax.sum()
+
+    #     norch_result.backward()
+    #     norch_tensor_grad = utils.to_torch(norch_tensor.grad).to(self.device)
+
+    #     torch_tensor = torch.tensor([[[10, 1], [-4, 0]], [[5., 50], [7, 8]]], requires_grad=True).to(self.device)
+
+    #     torch_softmax = torch.softmax(torch_tensor, dim=2)
+    #     torch_result = torch_softmax.sum()
+    #     torch_result.backward()
+    #     torch_tensor_grad = torch_tensor.grad
+       
+    #     self.assertTrue(utils.compare_torch(norch_tensor_grad, torch_tensor_grad))
+
     
     def test_reshape(self):
         """
@@ -361,6 +824,115 @@ class TestTensorAutograd(unittest.TestCase):
         self.assertTrue(utils.compare_torch(norch_tensor_grad_reshape_matmul1, torch_tensor_grad_reshape_matmul1))
         self.assertTrue(utils.compare_torch(norch_tensor_grad_reshape_matmul2, torch_tensor_grad_reshape_matmul2))
 
+    def test_unsqueeze(self):
+        """
+        Test autograd from unsqueezing a tensor: tensor.unsqueeze(dim)
+        """
+        
+        # Unsqueeze at dim=0
+        norch_tensor_unsqueeze = norch.Tensor([[1., 2], [3, 4]], requires_grad=True).to(self.device)
+        norch_result_unsqueeze_0 = norch_tensor_unsqueeze.unsqueeze(0).sum()
+        norch_result_unsqueeze_0.backward()
+        norch_tensor_grad_unsqueeze_0 = utils.to_torch(norch_tensor_unsqueeze.grad).to(self.device)
+
+        torch_tensor_unsqueeze = torch.tensor([[1., 2], [3, 4]], requires_grad=True).to(self.device)
+        torch_result_unsqueeze_0 = torch_tensor_unsqueeze.unsqueeze(0).sum()
+        torch_result_unsqueeze_0.backward()
+        torch_tensor_grad_unsqueeze_0 = torch_tensor_unsqueeze.grad
+
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_unsqueeze_0, torch_tensor_grad_unsqueeze_0))
+
+        # Unsqueeze at dim=1
+        norch_tensor_unsqueeze = norch.Tensor([[1., 2.], [3, 4]], requires_grad=True).to(self.device)
+        norch_result_unsqueeze_1 = norch_tensor_unsqueeze.unsqueeze(1).sum()
+        norch_result_unsqueeze_1.backward()
+        norch_tensor_grad_unsqueeze_1 = utils.to_torch(norch_tensor_unsqueeze.grad).to(self.device)
+
+        torch_tensor_unsqueeze = torch.tensor([[1., 2.], [3, 4]], requires_grad=True).to(self.device)
+        torch_result_unsqueeze_1 = torch_tensor_unsqueeze.unsqueeze(1).sum()
+        torch_result_unsqueeze_1.backward()
+        torch_tensor_grad_unsqueeze_1 = torch_tensor_unsqueeze.grad
+
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_unsqueeze_1, torch_tensor_grad_unsqueeze_1))
+
+        # Unsqueeze at dim=2
+        norch_tensor_unsqueeze = norch.Tensor([[1., 2], [3, 4]], requires_grad=True).to(self.device)
+        norch_result_unsqueeze_2 = norch_tensor_unsqueeze.unsqueeze(2).sum()
+        norch_result_unsqueeze_2.backward()
+        norch_tensor_grad_unsqueeze_2 = utils.to_torch(norch_tensor_unsqueeze.grad).to(self.device)
+
+        torch_tensor_unsqueeze = torch.tensor([[1., 2], [3, 4]], requires_grad=True).to(self.device)
+        torch_result_unsqueeze_2 = torch_tensor_unsqueeze.unsqueeze(2).sum()
+        torch_result_unsqueeze_2.backward()
+        torch_tensor_grad_unsqueeze_2 = torch_tensor_unsqueeze.grad
+
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_unsqueeze_2, torch_tensor_grad_unsqueeze_2))
+
+    def test_unsqueeze_then_matmul(self):
+        """
+        Test autograd from unsqueezing a tensor then performing matrix multiplication: matmul(tensor1.unsqueeze(dim), tensor2)
+        """
+        norch_tensor1 = norch.Tensor([[1, 2], [3, 4]], requires_grad=True).to(self.device)
+        norch_tensor2 = norch.Tensor([[1, 2], [3, 4]], requires_grad=True).to(self.device)
+        
+        # Unsqueeze at dim=0 then matmul
+        norch_result_unsqueeze_matmul = (norch_tensor1 @ norch_tensor2.unsqueeze(0)).sum()
+        norch_result_unsqueeze_matmul.backward()
+        norch_tensor_grad_unsqueeze_matmul1 = utils.to_torch(norch_tensor1.grad).to(self.device)
+        norch_tensor_grad_unsqueeze_matmul2 = utils.to_torch(norch_tensor2.grad).to(self.device)
+
+        torch_tensor1 = torch.tensor([[1, 2], [3, 4]], dtype=torch.float32, requires_grad=True).to(self.device)
+        torch_tensor2 = torch.tensor([[1, 2], [3, 4]], dtype=torch.float32, requires_grad=True).to(self.device)
+        
+        torch_result_unsqueeze_matmul = (torch_tensor1 @ torch_tensor2.unsqueeze(0)).sum()
+        torch_result_unsqueeze_matmul.backward()
+        torch_tensor_grad_unsqueeze_matmul1 = torch_tensor1.grad
+        torch_tensor_grad_unsqueeze_matmul2 = torch_tensor2.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_unsqueeze_matmul1, torch_tensor_grad_unsqueeze_matmul1))
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_unsqueeze_matmul2, torch_tensor_grad_unsqueeze_matmul2))
+
+    def test_squeeze(self):
+        """
+        Test autograd from squeezing a tensor: tensor.squeeze(dim)
+        """
+        # Squeeze at dim=0
+        norch_tensor_squeeze = norch.Tensor([[[1., 2], [3, 4]]], requires_grad=True).to(self.device)
+        norch_result_squeeze_0 = norch_tensor_squeeze.squeeze(0).sum()
+        norch_result_squeeze_0.backward()
+        norch_tensor_grad_squeeze_0 = utils.to_torch(norch_tensor_squeeze.grad).to(self.device)
+
+        torch_tensor_squeeze = torch.tensor([[[1., 2], [3, 4]]], requires_grad=True).to(self.device)
+        torch_result_squeeze_0 = torch_tensor_squeeze.squeeze(0).sum()
+        torch_result_squeeze_0.backward()
+        torch_tensor_grad_squeeze_0 = torch_tensor_squeeze.grad
+
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_squeeze_0, torch_tensor_grad_squeeze_0))
+
+    def test_squeeze_then_matmul(self):
+        """
+        Test autograd from squeezing a tensor then performing matrix multiplication: matmul(tensor1.squeeze(dim), tensor2)
+        """
+        norch_tensor1 = norch.Tensor([[[1., 2], [3, 4]]], requires_grad=True).to(self.device)
+        norch_tensor2 = norch.Tensor([[[1., 2], [3, 4]]], requires_grad=True).to(self.device)
+        
+        # Squeeze at dim=0 then matmul
+        norch_result_squeeze_matmul = (norch_tensor1.squeeze(0) @ norch_tensor2).sum()
+        norch_result_squeeze_matmul.backward()
+        norch_tensor_grad_squeeze_matmul1 = utils.to_torch(norch_tensor1.grad).to(self.device)
+        norch_tensor_grad_squeeze_matmul2 = utils.to_torch(norch_tensor2.grad).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[1., 2], [3, 4]]], dtype=torch.float32, requires_grad=True).to(self.device)
+        torch_tensor2 = torch.tensor([[[1., 2], [3, 4]]], dtype=torch.float32, requires_grad=True).to(self.device)
+        
+        torch_result_squeeze_matmul = (torch_tensor1.squeeze(0) @ torch_tensor2).sum()
+        torch_result_squeeze_matmul.backward()
+        torch_tensor_grad_squeeze_matmul1 = torch_tensor1.grad
+        torch_tensor_grad_squeeze_matmul2 = torch_tensor2.grad
+        
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_squeeze_matmul1, torch_tensor_grad_squeeze_matmul1))
+        self.assertTrue(utils.compare_torch(norch_tensor_grad_squeeze_matmul2, torch_tensor_grad_squeeze_matmul2))
+
 
     def test_T_then_matmul(self):
         """
diff --git a/tests/test_nn.py b/tests/test_nn.py
index e790fa9..cd623f3 100644
--- a/tests/test_nn.py
+++ b/tests/test_nn.py
@@ -21,13 +21,13 @@ class TestNNModuleLoss(unittest.TestCase):
         loss_fn_torch = torch.nn.MSELoss()
 
         # Test case 1: Predictions and labels are equal
-        predictions_norch = norch.Tensor([1.1, 2, 3, 4]).to(self.device)
-        labels_norch = norch.Tensor([1.1, 2, 3, 4]).to(self.device)
+        predictions_norch = norch.Tensor([[1.1, 2, 3, 4], [1.1, 2, 3, 4]]).to(self.device)
+        labels_norch = norch.Tensor([[1.1, 2, 3, 4], [1.1, 2, 3, 3]]).to(self.device)
         loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
         loss_torch_result = utils.to_torch(loss_norch).to(self.device)
 
-        predictions_torch = torch.tensor([1.1, 2, 3, 4]).to(self.device)
-        labels_torch = torch.tensor([1.1, 2, 3, 4]).to(self.device)
+        predictions_torch = torch.tensor([[1.1, 2, 3, 4], [1.1, 2, 3, 4]]).to(self.device)
+        labels_torch = torch.tensor([[1.1, 2, 3, 4], [1.1, 2, 3, 3]]).to(self.device)
         loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)        
         
         self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
@@ -44,6 +44,75 @@ class TestNNModuleLoss(unittest.TestCase):
         
         self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
 
+    def test_cross_entropy_loss(self):
+        """
+        Test the CrossEntropyLoss
+        """
+        loss_fn_norch = norch.nn.CrossEntropyLoss()
+        loss_fn_torch = torch.nn.CrossEntropyLoss()
+
+        # Test case 1: Single class, single sample
+        predictions_norch = norch.Tensor([2.0, 1.0, 0.1]).to(self.device)
+        labels_norch = norch.Tensor([0]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+
+        loss_torch_result = utils.to_torch(loss_norch).to(self.device)
+
+        predictions_torch = torch.tensor([2.0, 1.0, 0.1]).to(self.device)
+        labels_torch = torch.tensor(0).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+
+        self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
+
+        # Test case 2: Multiple classes, multiple samples
+        predictions_norch = norch.Tensor([[0.5, 1.5, 2.5], [1.0, 2.0, 3.0]]).to(self.device)
+        labels_norch = norch.Tensor([2, 1]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_torch_result = utils.to_torch(loss_norch).to(self.device)
+        
+
+        predictions_torch = torch.tensor([[0.5, 1.5, 2.5], [1.0, 2.0, 3.0]]).to(self.device)
+        labels_torch = torch.tensor([2, 1]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        
+        self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
+        
+        # Test case 3: Edge case - all predictions are zero
+        predictions_norch = norch.Tensor([[0.0, 0.0, 0.0], [0.0, 0.0, 0.0]]).to(self.device)
+        labels_norch = norch.Tensor([1, 2]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_torch_result = utils.to_torch(loss_norch).to(self.device)
+        
+        predictions_torch = torch.tensor([[0.0, 0.0, 0.0], [0.0, 0.0, 0.0]]).to(self.device)
+        labels_torch = torch.tensor([1, 2]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        
+        self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
+
+        # Test case 4: Class probabilities instead of class index
+        predictions_norch = norch.Tensor([0.5, 0.2, 0.1]).to(self.device)
+        labels_norch = norch.Tensor([1., 0, 0]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_torch_result = utils.to_torch(loss_norch).to(self.device)
+        
+        predictions_torch = torch.tensor([0.5, 0.2, 0.1]).to(self.device)
+        labels_torch = torch.tensor([1., 0, 0]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        
+        self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
+
+        # Test case 4: Batched class probabilities instead of class index
+        predictions_norch = norch.Tensor([[0.5, 0.2, 0.1], [0.1, 0.5, 0.7]]).to(self.device)
+        labels_norch = norch.Tensor([[1., 0, 0], [0, 1, 0]]).to(self.device)
+        loss_norch = loss_fn_norch.forward(predictions_norch, labels_norch)
+        loss_torch_result = utils.to_torch(loss_norch).to(self.device)
+
+        predictions_torch = torch.tensor([[0.5, 0.2, 0.1], [0.1, 0.5, 0.7]]).to(self.device)
+        labels_torch = torch.tensor([[1., 0, 0], [0, 1, 0]]).to(self.device)
+        loss_torch_expected = loss_fn_torch(predictions_torch, labels_torch)
+        
+        self.assertTrue(utils.compare_torch(loss_torch_result, loss_torch_expected))
+
 
 class TestNNModuleActivationFn(unittest.TestCase):
 
@@ -60,11 +129,11 @@ class TestNNModuleActivationFn(unittest.TestCase):
         sigmoid_fn_torch = torch.nn.Sigmoid()
 
         # Test case 1: Positive input
-        x = norch.Tensor([1, 2, 3]).to(self.device)
+        x = norch.Tensor([[1, 2, 3]]).to(self.device)
         sigmoid_norch = sigmoid_fn_norch.forward(x)
         sigmoid_torch_result = utils.to_torch(sigmoid_norch).to(self.device)
 
-        x = torch.tensor([1, 2, 3]).to(self.device)
+        x = torch.tensor([[1, 2, 3]]).to(self.device)
         sigmoid_torch_expected = sigmoid_fn_torch.forward(x)
 
         self.assertTrue(utils.compare_torch(sigmoid_torch_result, sigmoid_torch_expected))
@@ -87,4 +156,39 @@ class TestNNModuleActivationFn(unittest.TestCase):
         x = torch.tensor([0, 0, 0]).to(self.device)
         sigmoid_torch_expected = sigmoid_fn_torch.forward(x)
 
-        self.assertTrue(utils.compare_torch(sigmoid_torch_result, sigmoid_torch_expected))
\ No newline at end of file
+        self.assertTrue(utils.compare_torch(sigmoid_torch_result, sigmoid_torch_expected))
+
+    def test_softmax_activation(self):
+        """
+        Test Softmax activation function
+        """    
+
+        # Test different axes
+        axes = [0, 1, 2, -1]
+
+        # Define the input tensors for different test cases
+        test_cases = [
+            (norch.Tensor([[[1., 2, 3], [4, 5, 6]]]), torch.tensor([[[1., 2, 3], [4, 5, 6]]])),
+            (norch.Tensor([[[1., -1, 0], [2, -2, 0]]]), torch.tensor([[[1., -1, 0], [2, -2, 0]]])),
+            (norch.Tensor([[[0., 0, 0], [0, 0, 0]]]), torch.tensor([[[0., 0, 0], [0, 0, 0]]]))
+        ]
+
+        for dim in axes:
+            softmax_fn_norch = norch.nn.Softmax(dim=dim)
+            softmax_fn_torch = torch.nn.Softmax(dim=dim)
+
+            for norch_input, torch_input in test_cases:
+                # Move tensors to the correct device
+                norch_input = norch_input.to(self.device)
+                torch_input = torch_input.to(self.device)
+
+                # Forward pass using norch
+                softmax_norch = softmax_fn_norch.forward(norch_input)
+                softmax_torch_result = utils.to_torch(softmax_norch).to(self.device)
+
+                # Forward pass using torch
+                softmax_torch_expected = softmax_fn_torch.forward(torch_input)
+
+                # Compare the results
+                self.assertTrue(utils.compare_torch(softmax_torch_result, softmax_torch_expected))
+
diff --git a/tests/test_operations.py b/tests/test_operations.py
index 449c531..868f982 100644
--- a/tests/test_operations.py
+++ b/tests/test_operations.py
@@ -37,7 +37,7 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
-    def test_broadcasting_addition(self):
+    def test_addition_broadcasted(self):
         """
         Test addition of two tensors with broadcasting: tensor1 + tensor2
         """
@@ -52,6 +52,29 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+        norch_tensor1 = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]]).to(self.device)  # Shape (1, 2, 3)
+        norch_tensor2 = norch.Tensor([[10, 10], [5, 6]]).to(self.device)  # Shape (3)
+        norch_result = norch_tensor1 + norch_tensor2
+        torch_result = utils.to_torch(norch_result).to(self.device)
+        
+        torch_tensor1 = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]]).to(self.device)  # Shape (1, 2, 3)
+        torch_tensor2 = torch.tensor([[[10, 10], [5, 6]]]).to(self.device)  # Shape (3)
+        torch_expected = torch_tensor1 + torch_tensor2
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+        # reversed order broadcasting
+        norch_tensor1 = norch.Tensor([[0, 2]]).to(self.device) 
+        norch_tensor2 = norch.Tensor([[3, 4], [5, -1]]).to(self.device) 
+        norch_result = norch_tensor1 + norch_tensor2
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor1 = torch.tensor([[0, 2]]).to(self.device)  
+        torch_tensor2 = torch.tensor([[3, 4], [5, -1]]).to(self.device) 
+        torch_expected = torch_tensor1 + torch_tensor2
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
         norch_result = norch_tensor2 + norch_tensor1
         torch_expected = torch_tensor2 + torch_tensor1
 
@@ -89,6 +112,14 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+        # reversed order broadcasting
+        norch_result = norch_tensor2 - norch_tensor1
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_expected = torch_tensor2 - torch_tensor1
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
 
     def test_division_by_scalar(self):
         """
@@ -176,6 +207,94 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+    def test_unsqueeze(self):
+        """
+        Test unsqueeze operation on a tensor
+        """
+        norch_tensor = norch.Tensor([[1, 2], [3, 4]]).to(self.device)
+        
+        # Unsqueeze at dim=0
+        norch_unsqueeze_0 = norch_tensor.unsqueeze(0)
+        torch_unsqueeze_0 = utils.to_torch(norch_unsqueeze_0).to(self.device)
+        torch_tensor = torch.tensor([[1, 2], [3, 4]]).to(self.device)
+        torch_expected_0 = torch_tensor.unsqueeze(0)
+        self.assertTrue(utils.compare_torch(torch_unsqueeze_0, torch_expected_0))
+
+        # Unsqueeze at dim=1
+        norch_unsqueeze_1 = norch_tensor.unsqueeze(1)
+        torch_unsqueeze_1 = utils.to_torch(norch_unsqueeze_1).to(self.device)
+        torch_expected_1 = torch_tensor.unsqueeze(1)
+        self.assertTrue(utils.compare_torch(torch_unsqueeze_1, torch_expected_1))
+
+        # Unsqueeze at dim=2
+        norch_unsqueeze_2 = norch_tensor.unsqueeze(2)
+        torch_unsqueeze_2 = utils.to_torch(norch_unsqueeze_2).to(self.device)
+        torch_expected_2 = torch_tensor.unsqueeze(2)
+        self.assertTrue(utils.compare_torch(torch_unsqueeze_2, torch_expected_2))
+
+        # Unsqueeze at dim=-1
+        norch_unsqueeze_neg_1 = norch_tensor.unsqueeze(-1)
+        torch_unsqueeze_neg_1 = utils.to_torch(norch_unsqueeze_neg_1).to(self.device)
+        torch_expected_neg_1 = torch_tensor.unsqueeze(-1)
+        self.assertTrue(utils.compare_torch(torch_unsqueeze_neg_1, torch_expected_neg_1))
+
+        # Unsqueeze at dim=-2
+        norch_unsqueeze_neg_2 = norch_tensor.unsqueeze(-2)
+        torch_unsqueeze_neg_2 = utils.to_torch(norch_unsqueeze_neg_2).to(self.device)
+        torch_expected_neg_2 = torch_tensor.unsqueeze(-2)
+        self.assertTrue(utils.compare_torch(torch_unsqueeze_neg_2, torch_expected_neg_2))
+
+    def test_squeeze(self):
+        """
+        Test squeeze operation on a tensor
+        """
+        # Create a tensor with some dimensions of size 1
+        norch_tensor = norch.Tensor([[[1, 2], [3, 4]]]).to(self.device)  # shape [1, 2, 2]
+        
+        # Squeeze at dim=0
+        norch_squeeze_0 = norch_tensor.squeeze(0)
+        torch_squeeze_0 = utils.to_torch(norch_squeeze_0).to(self.device)
+        torch_tensor = torch.tensor([[[1, 2], [3, 4]]]).to(self.device)
+        torch_expected_0 = torch_tensor.squeeze(0)
+        self.assertTrue(utils.compare_torch(torch_squeeze_0, torch_expected_0))
+
+        # Squeeze at dim=2 (should raise an error because size is not 1)
+        with self.assertRaises(ValueError):
+            norch_tensor.squeeze(2)
+
+        # Create a tensor with a dimension of size 1 in the middle
+        norch_tensor_middle_1 = norch.Tensor([[[1, 2]], [[3, 4]]]).to(self.device)  # shape [2, 1, 2]
+        
+        # Squeeze at dim=1
+        norch_squeeze_1 = norch_tensor_middle_1.squeeze(1)
+        torch_squeeze_1 = utils.to_torch(norch_squeeze_1).to(self.device)
+        torch_tensor_middle_1 = torch.tensor([[[1, 2]], [[3, 4]]]).to(self.device)
+        torch_expected_1 = torch_tensor_middle_1.squeeze(1)
+        self.assertTrue(utils.compare_torch(torch_squeeze_1, torch_expected_1))
+
+        # Squeeze at dim=-2 (same as dim=1 in this case)
+        norch_squeeze_neg_2 = norch_tensor_middle_1.squeeze(-2)
+        torch_squeeze_neg_2 = utils.to_torch(norch_squeeze_neg_2).to(self.device)
+        torch_expected_neg_2 = torch_tensor_middle_1.squeeze(-2)
+        self.assertTrue(utils.compare_torch(torch_squeeze_neg_2, torch_expected_neg_2))
+
+        # Squeeze all dimensions of size 1 (None)
+        norch_tensor_all_1 = norch.Tensor([[[[1, 2], [3, 4]]]]).to(self.device)  # shape [1, 1, 2, 2]
+        norch_squeeze_all = norch_tensor_all_1.squeeze()
+        torch_squeeze_all = utils.to_torch(norch_squeeze_all).to(self.device)
+        torch_tensor_all_1 = torch.tensor([[[[1, 2], [3, 4]]]]).to(self.device)
+        torch_expected_all = torch_tensor_all_1.squeeze()
+        self.assertTrue(utils.compare_torch(torch_squeeze_all, torch_expected_all))
+
+        # Squeeze no dimensions (no dimensions of size 1)
+        norch_tensor_no_1 = norch.Tensor([[1, 2], [3, 4]]).to(self.device)  # shape [2, 2]
+        norch_squeeze_none = norch_tensor_no_1.squeeze()
+        torch_squeeze_none = utils.to_torch(norch_squeeze_none).to(self.device)
+        torch_tensor_no_1 = torch.tensor([[1, 2], [3, 4]]).to(self.device)
+        torch_expected_none = torch_tensor_no_1.squeeze()
+        self.assertTrue(utils.compare_torch(torch_squeeze_none, torch_expected_none))
+
+
     def test_transpose(self):
         """
         Test transposition of a tensor: tensor.transpose(dim1, dim2)
@@ -229,6 +348,134 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+        # negative axis
+
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.sum(axis=-2)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected = torch.sum(torch_tensor, dim=-2)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+
+    def test_sum_axis_keepdim(self):
+        """
+        Test summation of a tensor along a specific axis with keepdim=True
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.sum(axis=1, keepdim=True)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected = torch.sum(torch_tensor, dim=1, keepdim=True)
+        
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+    def test_max(self):
+        """
+        Test max of a tensor: tensor.max()
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.max()
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected = torch.max(torch_tensor)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+    def test_max_axis(self):
+        """
+        Test max of a tensor along a specific axis without keeping the dimensions
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.max(axis=1)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected, _ = torch.max(torch_tensor, dim=1)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+        # negative axis
+
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.max(axis=-1)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected, _ = torch.max(torch_tensor, dim=-1)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+
+    def test_max_axis_keepdim(self):
+        """
+        Test max of a tensor along a specific axis with keepdim=True
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.max(axis=1, keepdim=True)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected, _ = torch.max(torch_tensor, dim=1, keepdim=True)
+        
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+    def test_min(self):
+        """
+        Test min of a tensor: tensor.min()
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.min()
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected = torch.min(torch_tensor)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+    def test_min_axis(self):
+        """
+        Test min of a tensor along a specific axis without keeping the dimensions
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.min(axis=1)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected, _ = torch.min(torch_tensor, dim=1)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+        # negative axis
+
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.min(axis=-1)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected, _ = torch.min(torch_tensor, dim=-1)
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+
+    def test_min_axis_keepdim(self):
+        """
+        Test min of a tensor along a specific axis with keepdim=True
+        """
+        norch_tensor = norch.Tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        norch_result = norch_tensor.min(axis=1, keepdim=True)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor = torch.tensor([[[1, 2], [3, -4]], [[5, 6], [7, 8]]]).to(self.device)
+        torch_expected, _ = torch.min(torch_tensor, dim=1, keepdim=True)
+        
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+
     def test_transpose_T(self):
         """
         Test transposition of a tensor: tensor.T
@@ -242,6 +489,26 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+    def test_matmul(self):
+        """
+        Test matrix multiplication: MxP = NxM @ MxP
+        """
+        # Creating batched tensors for Norch
+        norch_tensor1 = norch.Tensor([[1, 2], [3, -4], [5, 6], [7, 8]]).to(self.device)
+        norch_tensor2 = norch.Tensor([[2, 3, 1, 0, 4], [5, -1, 2, 3, 0]]).to(self.device)
+
+        norch_result = norch_tensor1 @ norch_tensor2
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        # Converting to PyTorch tensors for comparison
+        torch_tensor1 = torch.tensor([[1, 2], [3, -4], [5, 6], [7, 8]]).to(self.device)
+        torch_tensor2 = torch.tensor([[2, 3, 1, 0, 4], [5, -1, 2, 3, 0]]).to(self.device)
+
+        torch_expected = torch.matmul(torch_tensor1, torch_tensor2)
+
+        # Comparing results
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
     def test_reshape_then_matmul(self):
         """
         Test reshaping a tensor followed by matrix multiplication: (tensor.reshape(shape) @ other_tensor)
@@ -258,6 +525,52 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+    def test_batched_matmul(self):
+        """
+        Test batched matrix multiplication: BxMxP = BxNxM @ BxMxP
+        """
+        B = 3  # Batch size
+
+        # Creating batched tensors for Norch
+        norch_tensor1 = norch.Tensor([[[1, 2], [3, -4], [5, 6], [7, 8]] for _ in range(B)]).to(self.device)
+        norch_tensor2 = norch.Tensor([[[2, 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)]).to(self.device)
+
+        norch_result = norch_tensor1 @ norch_tensor2
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        # Converting to PyTorch tensors for comparison
+        torch_tensor1 = torch.tensor([[[1, 2], [3, -4], [5, 6], [7, 8]] for _ in range(B)]).to(self.device)
+        torch_tensor2 = torch.tensor([[[2, 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)]).to(self.device)
+
+        torch_expected = torch.matmul(torch_tensor1, torch_tensor2)
+
+        # Comparing results
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+
+    def test_broadcasted_batched_matmul(self):
+        """
+        Test broadcasted batched matrix multiplication: BxMxP = NxM @ BxMxP
+        """
+        B = 3  # Batch size
+
+        # Creating batched tensors for Norch
+        norch_tensor1 = norch.Tensor([[1, 2], [3, -4], [5, 6], [7, 8]]).to(self.device)
+        norch_tensor2 = norch.Tensor([[[2, 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)]).to(self.device)
+
+        norch_result = norch_tensor1 @ norch_tensor2
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        # Converting to PyTorch tensors for comparison
+        torch_tensor1 = torch.tensor([[1, 2], [3, -4], [5, 6], [7, 8]]).to(self.device)
+        torch_tensor2 = torch.tensor([[[2, 3, 1, 0, 4], [5, -1, 2, 3, 0]] for _ in range(B)]).to(self.device)
+
+        torch_expected = torch.matmul(torch_tensor1, torch_tensor2)
+
+        # Comparing results
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+
 
     def test_transpose_then_matmul(self):
         """
@@ -346,6 +659,47 @@ class TestTensorOperations(unittest.TestCase):
 
         self.assertTrue(utils.compare_torch(torch_result, torch_expected))
 
+    def test_equal(self):
+        """
+        Test equal two tensors: tensor1.equal(tensor2)
+        """
+        norch_tensor1 = norch.Tensor([[[1, 2], [3, -4]], [[5, 1], [7, 8]]]).to(self.device)
+        norch_tensor2 = norch.Tensor([[[1, 1], [1, 1]], [[1, 1], [1, 1]]]).to(self.device)
+        norch_result = norch_tensor1.equal(norch_tensor2)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[1, 2], [3, -4]], [[5, 1], [7, 8]]]).to(self.device)
+        torch_tensor2 = torch.tensor([[[1, 1], [1, 1]], [[1, 1], [1, 1]]]).to(self.device)
+        torch_expected = (torch_tensor1 == torch_tensor2).float()
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+    def test_broadcasted_equal(self):
+        """
+        Test broadcasted equal two tensors: tensor1.equal(tensor2)
+        """
+        norch_tensor1 = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]]).to(self.device)
+        norch_tensor2 = norch.Tensor([[[10, 10]], [[5, 6]]]).to(self.device)
+        norch_result = norch_tensor1.equal(norch_tensor2)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]]).to(self.device)
+        torch_tensor2 = torch.tensor([[[10, 10]], [[5, 6]]]).to(self.device)
+        torch_expected = (torch_tensor1 == torch_tensor2).float()
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+
+        norch_tensor1 = norch.Tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]]).to(self.device)
+        norch_tensor2 = norch.Tensor([[[10.0,], [-4.0,]],[[6.0,], [8.0,]]]).to(self.device)
+        norch_result = norch_tensor1.equal(norch_tensor2)
+        torch_result = utils.to_torch(norch_result).to(self.device)
+
+        torch_tensor1 = torch.tensor([[[10, 10], [-4, -4]], [[5., 6], [7, 8]]]).to(self.device)
+        torch_tensor2 = torch.tensor([[[10.0,], [-4.0,]],[[6.0,], [8.0,]]]).to(self.device)
+        torch_expected = (torch_tensor1 == torch_tensor2).float()
+
+        self.assertTrue(utils.compare_torch(torch_result, torch_expected))
+    
 
     def test_zeros_like(self):
         """