modelscope / modelscope/DiffSynth-Studio
QwenImage多卡推理，单卡推理特别慢 正常吗
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 13.1k
- Forks
- 1.3k
- Avg merge
- 13h 12m
- Merged PRs (30d)
- 45
Description
运行的demo，模型本地化部署，单卡运行特别慢，3h还没出结果，一直卡在这
Image/vae/diffusion_pytorch_model.safetensors".
No qwen_image_blockwise_controlnet models available. This is not an error.
No siglip2_image_encoder models available. This is not an error.
No dinov3_image_encoder models available. This is not an error.
No qwen_image_image2lora_style models available. This is not an error.
No qwen_image_image2lora_coarse models available. This is not an error.
No qwen_image_image2lora_fine models available. This is not an error.
怎么多卡推理 加上usp参数 报 keyerror，wanpipeline 有这个参数 可以多卡推理
Contributor guide
No contributing guide indexed for this repository
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
Start by reproducing the local QwenImage demo and recording where single-card inference remains after loading Image/vae/diffusion_pytorch_model.safetensors. Then inspect the WanPipeline multi-card path and the USP argument handling to reproduce the reported KeyError. Done means explaining the slowdown or correcting it, and making supported multi-card inference work without that error.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- 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