Skip to content

Fix mlflow logging by closing the trainer properly #133

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
irenedea merged 1 commit into from
Aug 5, 2025
Merged

Fix mlflow logging by closing the trainer properly #133

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: 8 additions & 0 deletions test_single_controller_ppo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -284,6 +284,10 @@ def build_ppo_trainer(self):
load_path=self.ref_path,
)

def close_trainer(self):
self.ppo_trainer.close()


def add_rollouts(self, current_rank_rollouts: dict[str, Any]):
"""Adds the current rank's rollouts to the callback."""
for k, v in current_rank_rollouts.items():
Expand Down Expand Up @@ -474,6 +478,7 @@ def get_next_iter_rollouts(self):

processed_sequences = torch.cat([all_prompts, padded_responses], dim=-1)
iter_data['sequences'] = processed_sequences

return iter_data


Expand Down Expand Up @@ -652,6 +657,9 @@ def train(self):
# Populate the train actor group with the rollouts and then train
self.train_actor.add_latest_rollouts_from_buffer(self.experience_buffer)
self.train_actor.collective_methods.train_1_iter()

self.train_actor.collective_methods.close_trainer()



def _run_single_controller_ppo(
Expand Down