huggingface / huggingface/diffusers
Context parallel bug when using QwenImage
- Dominant language
- Python
- Stars
- 34.5k
- Forks
- 7.3k
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 91
Description
Refer more here: https://github.com/huggingface/diffusers/pull/12443
More context under "Result: Single coherent image matching single-GPU output" [here](https://github.com/huggingface/diffusers/pull/12443#issue-3489494846).
Before proceeding with a PR, please reply to this issue mentioning the proposal you have in mind, and we can take it from there.
Please mention @DN6 @yiyixuxu and myself (@sayakpaul).
Contributor guide
Research direction
Read pull request #12443, especially the section titled “Result: Single coherent image matching single-GPU output,” to understand the context-parallel QwenImage problem and proposed direction. Before starting implementation, reply to the issue with your proposal and mention @DN6, @yiyixuxu, and @sayakpaul; done means matching the single-GPU output coherently.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- 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