Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Grad norm tensor fix #938

Merged
merged 3 commits into from
May 1, 2024
Merged

Grad norm tensor fix #938

Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
e66ed72
Fix grad_norm tensor issue when eager mode run
yeonsily Apr 30, 2024
d369256
Fix code style
yeonsily Apr 30, 2024
5535812
Fix
regisss May 1, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Fix
  • Loading branch information
@regisss
regisss committed May 1, 2024
commit 5535812e61b2390a6e8fb4dc0188ead72ba318f3
4 changes: 2 additions & 2 deletions optimum/habana/transformers/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1184,11 +1184,11 @@ def _maybe_log_save_evaluate(self, tr_loss, _grad_norm, model, trial, epoch, ign
grad_norm = model.get_global_grad_norm()
else:
if _grad_norm is not None and self.accelerator.distributed_type != GaudiDistributedType.FSDP:
grad_norm = _grad_norm.item() if _grad_norm.size() == 1 else _grad_norm.tolist()
grad_norm = _grad_norm.item() if _grad_norm.size() == torch.Size([1]) else _grad_norm.tolist()
else:
grad_norm = None

if grad_norm is not None:
if grad_norm is not None and not isinstance(grad_norm, list):
logs["grad_norm"] = grad_norm
logs["learning_rate"] = self._get_learning_rate()

Expand Down
Loading