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

[Trainer] Fix loading rng state #9656

Merged
merged 1 commit into from
Dec 19, 2024
+8 −3
Merged

[Trainer] Fix loading rng state #9656

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
11 changes: 8 additions & 3 deletions paddlenlp/trainer/trainer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1904,9 +1904,14 @@
if "hybrid_parallel_rng_state_tracker" in checkpoint_rng_state:
if self.args.tensor_parallel_degree <= 1:
checkpoint_rng_state["hybrid_parallel_rng_state_tracker"].pop("model_parallel_rng", None)
fleet.meta_parallel.get_rng_state_tracker().set_states_tracker(
checkpoint_rng_state["hybrid_parallel_rng_state_tracker"]
)
try:
fleet.meta_parallel.get_rng_state_tracker().set_states_tracker(

Check warning on line 1908 in paddlenlp/trainer/trainer.py

View check run for this annotation

Codecov / codecov/patch

paddlenlp/trainer/trainer.py#L1907-L1908 

Added lines #L1907 - L1908 were not covered by tests
checkpoint_rng_state["hybrid_parallel_rng_state_tracker"]
)
except:
logger.warning(

Check warning on line 1912 in paddlenlp/trainer/trainer.py

View check run for this annotation

Codecov / codecov/patch

paddlenlp/trainer/trainer.py#L1911-L1912 

Added lines #L1911 - L1912 were not covered by tests
"Hybrid paralell rng states change when training environment differs, so we dot not set state tracker here."
)
else:
logger.warning("Not found hybrid parallel RNG state.")

Expand Down