modelscope / modelscope/ms-swift
[RFC] Support for On-Policy Distillation with Student and Teacher from Different Model Families
Open
Nobody has claimed this yet.
enhancement
- Dominant language
- Python
- Stars
- 15.7k
- Forks
- 1.7k
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 136
Description
目前发现gkd的代码中是将student的inputs直接放入到teacher里面计算分布。如果student和teacher是不同family的模型则无法正常训练。
后续是否会支持对于student和teacher是不同family的On-Policy Distillation？
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
No files, tests, or entry points are named. Start by tracing the GKD on-policy distillation training flow where student inputs are passed to the teacher, then inspect how different model-family interfaces could be separated. Done should mean a supported training path for different student and teacher families, with coverage for that scenario.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- ai, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100