modelscope / modelscope/ms-swift

HunyuanOCR GRPO failed

Open
#7,159 7 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

stale
Dominant language
Python
Stars
15.7k
Forks
1.7k
Avg merge
1d 16h
Merged PRs (30d)
136

Description

Describe the bug
Thanks for your timely HunyuanOCR training support! I installed ms-swift from latest source, used your train.sh under examples/ and it worked fine. But when I changed sft \ to rlhf --rlhf_type grpo \, I got terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc. I tried my personal dataset, and the error persists.

Your hardware and system info
A800*8
CUDA 12.8
torch2.8
vllm nightly version
flash-attn 2.8.3
transformers 4.57.1.dev0

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 with the train.sh script under examples/ and reproduce the GRPO variant in the reported A800, CUDA, PyTorch, vLLM, flash-attn, and Transformers environment. Compare it with the working SFT invocation and trace the std::bad_alloc failure; done means identifying a reproducible cause and documenting or implementing a verified fix.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.