aparler/c-llm
1
0
Fork 0
mirror of https://github.com/karpathy/llm.c.git synced 2026-07-28 20:35:09 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix np.int32 overflow in token counting

Browse source
This commit is contained in:
Igor Shilov 2025-07-03 17:56:28 +00:00
parent f1e2ace651
commit f51ab0383d
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
train_gpt2.py
View file

@ -342,7 +342,7 @@ class DistributedDataLoader:
for fname in self.files:
shard_ntok = _peek_data_shard(fname)
assert shard_ntok >= num_processes * B * T + 1
ntok_total += shard_ntok
ntok_total += int(shard_ntok)
self.ntok_total = ntok_total
print0(f"DataLoader: total number of tokens: {ntok_total:,} across {len(self.files)} files")