Skip to content

Commit

Permalink
remove trace
Browse files Browse the repository at this point in the history
  • Loading branch information
Eugene-hu committed Aug 25, 2023
1 parent 00ea6fe commit 1a65224
Showing 1 changed file with 0 additions and 1 deletion.
1 change: 0 additions & 1 deletion openvalidators/reward/dpo.py
Original file line number Diff line number Diff line change
Expand Up @@ -107,7 +107,6 @@ def get_rewards(self, prompt: str, completions: List[str], name: str) -> torch.F


def logit_penalty(self, input_ids: torch.LongTensor, logit: torch.FloatTensor) -> torch.FloatTensor:
import pdb;pdb.set_trace()
score = torch.gather(logit, 1, input_ids.unsqueeze(0))

# if score < 0 then repetition penalty has to be multiplied to reduce the previous token probability
Expand Down

0 comments on commit 1a65224

Please sign in to comment.