modelscope / modelscope/ms-swift
ArenaRL support
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 15.7k
- Forks
- 1.7k
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 136
Description
Checklist / 检查清单
- I have searched existing issues, and this is a new feature request. / 我已经搜索过现有的 issues，确认这是一个新的 Feature Request。
Feature Request Description / Feature Request 描述
I would like to ask whether there are plans to support ArenaRL-style advantage computation, or if such functionality could be added in the future.This differs from standard reward-based RLHF in that:
- The baseline is per-prompt, not a learned value function or global running mean
- No pointwise reward scale consistency across prompts is required
Could you please let me know:
Whether the current framework already supports custom, per-group advantage injection?
If not, whether adding such support would align with the roadmap?
Best regards,
Pull Request / Pull Request 信息
No response
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
No files, tests, or entry points are named. Start by locating the framework's current advantage computation and determine whether custom per-group advantage injection is supported; done would require a clear implementation scope or maintainer decision about ArenaRL-style support.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- ai, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 28/100