Skip to content

fix optimizer number of steps #3452

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
copybara-service merged 1 commit into from
Mar 20, 2026
Merged

fix optimizer number of steps #3452

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
3 changes: 3 additions & 0 deletions src/maxtext/trainers/post_train/distillation/distillation_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -248,6 +248,9 @@ def compute_loss(
"distill/teacher_loss": teacher_hard_loss,
"distill/out_proj_feature_loss": feature_loss,
"distill/total_loss": total_loss,
"distill/temperature": self.temperature,
"distill/alpha": self.alpha,
"distill/beta_feature": self.beta_feature,
}
return total_loss, metrics

Expand Down
3 changes: 1 addition & 2 deletions src/maxtext/trainers/post_train/distillation/train_distill.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -464,8 +464,7 @@ def labels_fn(targets, targets_segmentation=None, **kwargs):
)

# 4. Optimizer & Config
total_updates = student_config.steps // student_config.gradient_accumulation_steps
optimizer = get_distillation_optimizer(student_config, total_updates)
optimizer = get_distillation_optimizer(student_config, student_config.steps)

checkpointing_options = checkpoint.CheckpointManagerOptions(
save_interval_steps=student_config.checkpoint_period,
Expand Down
3 changes: 3 additions & 0 deletions tests/post_training/unit/train_distill_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -399,6 +399,9 @@ def _test_monitored_strategy(self, sft_mode: bool):
"distill/teacher_loss",
"distill/out_proj_feature_loss",
"distill/total_loss",
"distill/temperature",
"distill/alpha",
"distill/beta_feature",
]
for key in expected_keys:
self.assertIn(key, metrics)
Expand Down
Loading