Skip to content

fix: forward dual-clip PPO epsilon - #2247

Open
alexqdh wants to merge 1 commit into
THUDM:mainfrom
alexqdh:add_eps_clip_c
Open

fix: forward dual-clip PPO epsilon#2247
alexqdh wants to merge 1 commit into
THUDM:mainfrom
alexqdh:add_eps_clip_c

Conversation

@alexqdh

@alexqdh alexqdh commented Jul 31, 2026

Copy link
Copy Markdown

Fix Dual-clip PPO support by forwarding args.eps_clip_c from policy_loss_function to compute_policy_loss. Forward eps_clip_c=args.eps_clip_c in the PPO policy-loss path.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant