Skip to content

fix(runners): guard against empty loss_history in all trainers #36

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
sacredvoid merged 1 commit into from
Mar 26, 2026
Merged

fix(runners): guard against empty loss_history in all trainers #36

Show file tree
Hide file tree
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: 2 additions & 0 deletions src/alignrl/dpo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -106,6 +106,8 @@ def train(self) -> TrainResult:
trainer.save_model(str(output_dir / "final"))

loss_history = [log["loss"] for log in trainer.state.log_history if "loss" in log]
if not loss_history:
loss_history = [result.training_loss]

train_result = TrainResult(
output_dir=output_dir / "final",
Expand Down
2 changes: 2 additions & 0 deletions src/alignrl/grpo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -135,6 +135,8 @@ def train(self) -> TrainResult:
trainer.save_model(str(output_dir / "final"))

loss_history = [log["loss"] for log in trainer.state.log_history if "loss" in log]
if not loss_history:
loss_history = [result.training_loss]
reward_history = [
log.get("reward", 0.0) for log in trainer.state.log_history if "reward" in log
]
Expand Down
2 changes: 2 additions & 0 deletions src/alignrl/sft.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,8 @@ def train(self) -> TrainResult:
trainer.save_model(str(output_dir / "final"))

loss_history = [log["loss"] for log in trainer.state.log_history if "loss" in log]
if not loss_history:
loss_history = [result.training_loss]

train_result = TrainResult(
output_dir=output_dir / "final",
Expand Down
Loading