modelscope / modelscope/DiffSynth-Studio
zero3保存模型通信错误
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 13.1k
- Forks
- 1.3k
- Avg merge
- 13h 12m
- Merged PRs (30d)
- 45
Description
sft训练Wan2.1-T2V-14B，开启zero3，在accelerate.yaml中设置了zero3_save_16bit_model: true，存储模型通信卡死，根据日志发现卡死原因，rank 0使用op allgather_base，其他rank使用op allreduce
换用zero2可以正常工作
这是deepspeed的bug吗，还是diffsynth模型需要适配？
Contributor guide
No contributing guide indexed for this repository
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
Start by reproducing the Wan2.1-T2V-14B SFT run with zero3_save_16bit_model: true in accelerate.yaml and inspect the reported logs. Compare the rank 0 allgather_base path with the other ranks' allreduce path, using the working zero2 configuration as a baseline. Done means identifying whether the deadlock is in DeepSpeed or requires a DiffSynth-Studio adaptation and documenting a reproducible resolution.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- distributed-systems, machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100