modelscope / modelscope/ms-swift
Lora qw3-vl in my VQA datasets, lora's acc is very high, but acc is too low in some new images.
Open
Nobody has claimed this yet.
question
stale
- Dominant language
- Python
- Stars
- 15.7k
- Forks
- 1.7k
- Avg merge
- 1d 16h
- Merged PRs (30d)
- 136
Description
Checklist / 检查清单
- I have searched existing issues, and this is a new question or discussion topic. / 我已经搜索过现有的 issues，确认这是一个新的问题与讨论。
Question Description / 问题描述
Lora qw3-vl in my VQA datasets, lora's token_acc is very high, but token_acc is too low in some new images.
- Lora my VQA datasets, the token_acc is very high, up to 99%.
2.But in new images, the same question, it's token_acc is 60%.
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 names no files, tests, or entry points. Start by reproducing the reported token_acc difference on the VQA training datasets and the new images, then compare the data and evaluation conditions. Done means identifying the cause of the accuracy gap and documenting or addressing it.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 32/100