THUDM / THUDM/slime

Error when training MoE with PPO

Open
#1,108 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
8.5k
Forks
1.3k
Avg merge
5h 36m
Merged PRs (30d)
22

Description

Hi, thanks for this great project. I got an error when training MoE (qwen30b-a3b) using PPO. I have tried GRPO, and it was all good. Now I changed --advantage-estimator to ppo, but got an error:

Image

It seems that the advantage is None. How can I fix this problem? And it would also help if you could provide an example script for using PPO, since currently most examples are using GRPO. Thank you!

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing MoE training with qwen30b-a3b and --advantage-estimator ppo, then compare the failing path with the working GRPO configuration. Trace where the advantage becomes None and document the required PPO setup; done means PPO training works and an example script is available.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.