github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-10-29 04:40:36 +00:00
Code Issues Packages Projects Releases Wiki Activity

[misc] update pre-commit and run all files (#4752)

Browse Source 
* [misc] update pre-commit

* [misc] run pre-commit

* [misc] remove useless configuration files

* [misc] ignore cuda for clang-format
This commit is contained in:
Hongxin Liu
2023-09-19 14:20:26 +08:00
committed by GitHub
parent 3c6b831c26
commit 079bf3cb26
1268 changed files with 50037 additions and 38444 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
examples/tutorial/sequence_parallel/loss_func/cross_entropy.py
View File

@@ -1,11 +1,8 @@
import torch
from torch.cuda.amp import custom_bwd, custom_fwd
from colossalai.legacy.context.parallel_mode import ParallelMode
class _VocabCrossEntropy(torch.autograd.Function):
@staticmethod
@custom_fwd
def forward(ctx, vocab_parallel_logits, target):
@@ -59,7 +56,7 @@ class _VocabCrossEntropy(torch.autograd.Function):
# Add the gradient from matching classes.
arange_1d = torch.arange(start=0, end=grad_2d.size()[0], device=grad_2d.device)
grad_2d[arange_1d, masked_target_1d] -= (1.0 - target_mask.view(-1).float())
grad_2d[arange_1d, masked_target_1d] -= 1.0 - target_mask.view(-1).float()
# Finally elementwise multiplication with the output gradients.
grad_input.mul_(grad_output.unsqueeze(dim=-1))