modelscope / modelscope/ms-swift
[讨论] 大家有遇到过sft开关deepspeed效果差异大的情况吗
Open
Nobody has claimed this yet.
good first issue
- Dominant language
- Python
- Stars
- 15.7k
- Forks
- 1.7k
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 136
Description
swift sft相同数据，训练脚本相差一个参数 --deepspeed zero2, 开zero2的指标高五个点。
怀疑是有些操作(gradient clip)实现不一样
大家有遇到这样的情况吗
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
Start by reproducing the reported swift SFT comparison with identical data and scripts, changing only the --deepspeed zero2 parameter. Compare the resulting metrics and investigate whether gradient clipping or another training operation differs; done means identifying and documenting the cause of the five-point discrepancy.
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
- 25/100