Skip to content

FSDP bug fix for load_state_dict #18596

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
merged 1 commit into from
Aug 12, 2022
Merged

FSDP bug fix for load_state_dict #18596

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
8 changes: 6 additions & 2 deletions src/transformers/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1935,7 +1935,9 @@ def _load_from_checkpoint(self, resume_from_checkpoint, model=None):
else:
# We load the model state dict on the CPU to avoid an OOM error.
state_dict = torch.load(os.path.join(resume_from_checkpoint, WEIGHTS_NAME), map_location="cpu")
load_result = model.load_state_dict(state_dict, strict=False)
# workaround for FSDP bug https://github.com/pytorch/pytorch/issues/82963
# which takes *args instead of **kwargs
load_result = model.load_state_dict(state_dict, False)
# release memory
del state_dict
self._issue_warnings_after_load(load_result)
Expand Down Expand Up @@ -1989,7 +1991,9 @@ def _load_best_model(self):
# We load the model state dict on the CPU to avoid an OOM error.
state_dict = torch.load(best_model_path, map_location="cpu")
# If the model is on the GPU, it still works!
load_result = model.load_state_dict(state_dict, strict=False)
# workaround for FSDP bug https://github.com/pytorch/pytorch/issues/82963
# which takes *args instead of **kwargs
load_result = model.load_state_dict(state_dict, False)
if not is_sagemaker_mp_enabled():
self._issue_warnings_after_load(load_result)
elif os.path.exists(os.path.join(self.state.best_model_checkpoint, WEIGHTS_INDEX_NAME)):
Expand Down