Thanks to visit codestin.com
Credit goes to github.com

Skip to content

vlm grpo UnslothEfficientGRPO #188

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

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
Open

vlm grpo UnslothEfficientGRPO #188

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
50cba9a
Add logits hook post grad
GAD-cell Jun 16, 2025
c54e91c
grad post hook update
GAD-cell Jun 17, 2025
d2b41fe
Merge branch 'main' into VLM_GRPO
GAD-cell Jul 2, 2025
ba43cd4
added support for vision input in grpo_accumulated
GAD-cell Jul 2, 2025
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
Prev Previous commit
Next Next commit
grad post hook update
  • Loading branch information
@GAD-cell
GAD-cell committed Jun 17, 2025
commit c54e91cc2898ce0ed10d888e8f02a1dbbfe7d242
2 changes: 1 addition & 1 deletion unsloth_zoo/peft_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -197,7 +197,7 @@ def requires_grad_post_hook(module, input, output):
try: # For dataclass from HF, try on loss or logits
if hasattr(output, "loss") and output.loss is not None:
output.loss.requires_grad_(True)
elif hasattr(output, "logits") and output.logits is not None: #with RL like GRPO there are no loss as you don't provide labels
elif hasattr(output, "logits") and output.logits is not None: #with RL like GRPO there are no loss as you don't provide labels
output.logits.requires_grad_(True)
else:
raise ValueError("Neither loss nor logits are available for grad post hook.")
Expand Down