Skip to content

fix: correct margin calculation in DPO training #172

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
leejianwoo-collab wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: correct margin calculation in DPO training #172

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: 2 additions & 1 deletion tinker_cookbook/preference/train_dpo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -140,7 +140,7 @@ def compute_dpo_loss(
accuracy = (chosen_log_ratio > rejected_log_ratio).float().mean().item()
chosen_rewards = dpo_beta * chosen_log_ratio
rejected_rewards = dpo_beta * rejected_log_ratio
margin = dpo_beta * (chosen_rewards - rejected_rewards).mean().item()
margin = (chosen_rewards - rejected_rewards).mean().item()

metrics = {
"dpo_loss": loss.item(),
Expand Down Expand Up @@ -394,3 +394,4 @@ def print_example(datum: tinker.Datum, tokenizer: Tokenizer, label: str = ""):
weights = datum.loss_fn_inputs["weights"].data
logger.info(f"\n{label} Example:")
logger.info(format_colorized(int_tokens, cast(list[float], weights), tokenizer))