ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update rl_replacements.py

Browse source
This commit is contained in:
Daniel Han 2025-06-22 05:32:38 -07:00
commit 19d36655f7
1 changed files with 1 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

1
unsloth/models/rl_replacements.py
View file

@ -340,6 +340,7 @@ def grpo_trainer_compute_loss(function_name, function):
print("$$$$$$$$$$$ input_ids", input_ids.shape)
input_ids = input_ids[:, -logits_to_keep:]
print("$$$$$$$$$$$ input_ids", input_ids.shape)
if per_token_logps is not None:
if ref_per_token_logps is not None: