Skip to content

Fixing LLAMA-V2-70B-FT-HF (8x) issue for eager mode #35

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

Merged
vivekgoe merged 1 commit into from
Feb 13, 2024
Merged

Fixing LLAMA-V2-70B-FT-HF (8x) issue for eager mode #35

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion optimum/habana/transformers/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1425,7 +1425,7 @@ def training_step(self, model: torch.nn.Module, inputs: Dict[str, Union[torch.Te
if self.args.n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu parallel training

if self.args.pipelining_fwd_bwd:
if self.args.use_lazy_mode and self.args.pipelining_fwd_bwd:
self.htcore.mark_step()

self.accelerator.backward(loss)
Expand Down