modelscope / modelscope/ms-swift
为什么说在CPT/SFT训练中，梯度累加的训练效果等价使用更大的batch_size，但在RLHF训练中，训练效果并不等价。
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 15.7k
- Forks
- 1.7k
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 136
Description
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
The issue contains only an image and asks why gradient accumulation is equivalent to a larger batch size for CPT/SFT but not for RLHF. No files, tests, or entry points are identified, so first clarify the intended explanation and the relevant training paths before defining what a complete answer or documentation change would include.
Written by the indexing model from the issue text.
Assessment
- Domain
- machine-learning
- Issue type
- Documentation
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100