modelscope / modelscope/ms-swift
Data containing multiple images is not supported for the Qwen3.5 series.
Nobody has claimed this yet.
- 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 bug report. / 我已经搜索过现有的 issues，确认这是一个新的 bug report。
Bug Description / Bug 描述
When training on the Qwen 3.5 series, training proceeds normally if the multimodal dataset contains one image per sample. However, an OOM error occurs if the dataset contains more than one image per sample.
How to Reproduce / 如何复现
ms-swift: 4.4.0.dev0
torch=2.7.1
deepspeed=0.19.2 zero3
Additional Information / 补充信息
packing true
max_length 32768
image_max_token_num 1024
train on 32 pieces of 910B
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 source file or test is identified. First reproduce training with ms-swift 4.4.0.dev0, torch 2.7.1, DeepSpeed ZeRO-3, packing enabled, max_length 32768, and image_max_token_num 1024, comparing one image per sample with multiple images; done means the multi-image dataset trains without an OOM.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning, performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100