diff --git a/build/libtensor.so b/build/libtensor.so
index dd8d235..d7ec513 100755
Binary files a/build/libtensor.so and b/build/libtensor.so differ
diff --git a/csrc/tensor.cpp b/csrc/tensor.cpp
index 4b734ac..b3cfb3e 100644
--- a/csrc/tensor.cpp
+++ b/csrc/tensor.cpp
@@ -5,6 +5,8 @@
 extern "C" {
 
     Tensor* create_tensor(float* data, int* shape, int ndim) {
+
+        printf("Creating tensor\n");
         Tensor* tensor = (Tensor*)malloc(sizeof(Tensor));
         if (tensor == NULL) {
             fprintf(stderr, "Memory allocation failed\n");
@@ -51,12 +53,12 @@ extern "C" {
             }
         }
         printf("]\n\n\n");
-
         
         return tensor;
     }
 
     float get_item(Tensor* tensor, int* indices) {
+
         int index = 0;
         for (int i = 0; i < tensor->ndim; i++) {
             index += indices[i] * tensor->strides[i];
@@ -65,6 +67,7 @@ extern "C" {
     }
 
     Tensor* add_tensor(Tensor* tensor1, Tensor* tensor2) {
+        printf("Adding tensor\n");
         if (tensor1->ndim != tensor2->ndim) {
             fprintf(stderr, "Tensors must have the same number of dimensions %d and %d for addition\n", tensor1->ndim, tensor2->ndim);
             exit(1);
@@ -135,4 +138,77 @@ extern "C" {
 
         return create_tensor(result_data, shape, ndim);
     }
+
+    Tensor* sub_tensor(Tensor* tensor1, Tensor* tensor2) {
+        printf("Adding tensor\n");
+        if (tensor1->ndim != tensor2->ndim) {
+            fprintf(stderr, "Tensors must have the same number of dimensions %d and %d for addition\n", tensor1->ndim, tensor2->ndim);
+            exit(1);
+        }
+
+        int ndim = tensor1->ndim;
+        int* shape = (int*)malloc(ndim * sizeof(int));
+        if (shape == NULL) {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(1);
+        }
+
+        printf("Size: %d\n", tensor1->size);
+        printf("Data: [");
+        for (int i = 0; i < tensor1->size; i++) {
+            printf("%.2f", tensor1->data[i]);
+            if (i < tensor1->size - 1) {
+                printf(", ");
+            }
+        }
+        printf("]\n");
+
+        printf("Size: %d\n", tensor2->size);
+        printf("Data: [");
+        for (int i = 0; i < tensor2->size; i++) {
+            printf("%.2f", tensor2->data[i]);
+            if (i < tensor2->size - 1) {
+                printf(", ");
+            }
+        }
+        printf("]\n");
+
+        printf("Shapes : [");
+        for (int i = 0; i < tensor1->ndim; i++) {
+            printf("%d", tensor1->shape[i]);
+            if (i < tensor1->ndim - 1) {
+                printf(", ");
+            }
+        }
+        printf("]\n");
+
+        printf("Shapes : [");
+        for (int i = 0; i < tensor2->ndim; i++) {
+            printf("%d", tensor2->shape[i]);
+            if (i < tensor2->ndim - 1) {
+                printf(", ");
+            }
+        }
+        printf("]\n");
+
+        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 addition\n", tensor1->shape[i], tensor2->shape[i], i);
+                exit(1);
+            }
+            shape[i] = tensor1->shape[i];
+        }
+
+        float* result_data = (float*)malloc(tensor1->size * sizeof(float));
+        if (result_data == NULL) {
+            fprintf(stderr, "Memory allocation failed\n");
+            exit(1);
+        }
+
+        for (int i = 0; i < tensor1->size; i++) {
+            result_data[i] = tensor1->data[i] - tensor2->data[i];
+        }
+
+        return create_tensor(result_data, shape, ndim);
+    }
 }
diff --git a/csrc/tensor.h b/csrc/tensor.h
index 0ce9d86..58e79a2 100644
--- a/csrc/tensor.h
+++ b/csrc/tensor.h
@@ -14,6 +14,7 @@ extern "C" {
     float get_item(Tensor* tensor, int* indices);
     Tensor* create_tensor(float* data, int* shape, int ndim);
     Tensor* add_tensor(Tensor* tensor1, Tensor* tensor2);
+    Tensor* sub_tensor(Tensor* tensor1, Tensor* tensor2);
 }
 
 #endif /* TENSOR_H */
diff --git a/csrc/tensor.o b/csrc/tensor.o
index 0e5799c..363e2a1 100644
Binary files a/csrc/tensor.o and b/csrc/tensor.o differ
diff --git a/norch/__pycache__/tensor.cpython-38.pyc b/norch/__pycache__/tensor.cpython-38.pyc
index 8523f4f..d8a3609 100644
Binary files a/norch/__pycache__/tensor.cpython-38.pyc and b/norch/__pycache__/tensor.cpython-38.pyc differ
diff --git a/norch/tensor.py b/norch/tensor.py
index 8ac8d61..62ad203 100644
--- a/norch/tensor.py
+++ b/norch/tensor.py
@@ -12,21 +12,31 @@ class CTensor(ctypes.Structure):
 class Tensor:
     _C = ctypes.CDLL("../build/libtensor.so")
 
-    def __init__(self, data):
-        data, shape = self.flatten(data)
-          # Adjust the path to the shared library
-        self.data = (ctypes.c_float * len(data))(*data)
-        self.shape = shape
-        self.ndim = len(shape)
+    def __init__(self, data=None):
 
-        Tensor._C.create_tensor.argtypes = [ctypes.POINTER(ctypes.c_float), ctypes.POINTER(ctypes.c_int), ctypes.c_int]
-        Tensor._C.create_tensor.restype = ctypes.POINTER(CTensor)
+        if data != None:
+            data, shape = self.flatten(data)
+            # Adjust the path to the shared library
+            self.data_ctype = (ctypes.c_float * len(data))(*data)
+            self.shape_ctype = (ctypes.c_int * len(shape))(*shape)
+            self.ndim_ctype = ctypes.c_int(len(shape))
 
-        self.tensor = Tensor._C.create_tensor(
-            self.data,
-            (ctypes.c_int * len(shape))(*shape),
-            ctypes.c_int(len(shape))
-        )
+            self.shape = shape
+            self.ndim = len(shape)
+
+            Tensor._C.create_tensor.argtypes = [ctypes.POINTER(ctypes.c_float), ctypes.POINTER(ctypes.c_int), ctypes.c_int]
+            Tensor._C.create_tensor.restype = ctypes.POINTER(CTensor)
+
+            self.tensor = Tensor._C.create_tensor(
+                self.data_ctype,
+                self.shape_ctype,
+                self.ndim_ctype   
+            )
+        
+        else:
+            self.tensor = None,
+            self.shape = None,
+            self.ndim = None
 
     def flatten(self, nested_list):
         flat_data = []
@@ -68,11 +78,28 @@ class Tensor:
 
         result_tensor_ptr = Tensor._C.add_tensor(self.tensor, other.tensor)
 
-        result_data = [result_tensor_ptr.contents.data[i] for i in range(self.shape[0] * self.shape[1])]
-        result_shape = [result_tensor_ptr.contents.shape[i] for i in range(result_tensor_ptr.contents.ndim)]
+        result_data = Tensor()
+        result_data.tensor = result_tensor_ptr
+        result_data.shape = self.shape.copy()
+        result_data.ndim = self.ndim
 
-        return Tensor(result_data, result_shape)
+        return result_data
+    
+    def __sub__(self, other):
+        if self.shape != other.shape:
+            raise ValueError("Tensors must have the same shape for addition")
+        
+        Tensor._C.add_tensor.argtypes = [ctypes.POINTER(CTensor), ctypes.POINTER(CTensor)]
+        Tensor._C.add_tensor.restype = ctypes.POINTER(CTensor)
 
+        result_tensor_ptr = Tensor._C.sub_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
+
+        return result_data
 
 if __name__ == "__main__":
     from tensor import Tensor
@@ -82,7 +109,10 @@ if __name__ == "__main__":
     a = Tensor([[1, 2, 3], [1, 2, 3]])
     b = Tensor([[1, 2, 3], [1, 2, 3]])
 
-    c = a + b
+    c = a + b - b
+
+    print(c)
+    
     fim = time.time()
 
     print(fim-ini)
\ No newline at end of file