tutorials/image/qwen/qwen-image-edit

[giscus[bot]]

tutorials/image/qwen/qwen-image-edit

Qwen-Image-Edit is the image editing version of Qwen-Image, further trained based on the 20B model, supporting precise text editing and dual semantic/appearance editing capabilities.

https://docs.comfy.org/tutorials/image/qwen/qwen-image-edit

[tanglei807]

负面条件加上以后图像质量降低很厉害，为什么呢

[eoldrk]

Model does not work on MPS

[raffaelelu]

Hi community, I wonder how the model recognize the elements from additional images? eg if I want to replace a bottle in the input image with a bottle from image 2, how should I mention the secondary bottle? I tried "the bottle from image 2" but it doesn't work. the bottle is disappeared in the output image.

[CyberTodBG]

I have the exact same issue. The additional images are not recognized.
I see the images go into the CLIP node, but that's it.

[user1198]

Hi,
First of all, thank you so much for creating the Image Edit 2509 workflow—it has been one of my absolute favorites to use!
However, after recent ComfyUI updates, I’ve run into an issue where several models (like Qwen text/image edit models) have stopped working properly. Instead of generating the expected image edits, the final output just results in static/complex noise.
I wanted to reach out and bring this to your attention in case the recent ComfyUI core changes broke something in the nodes or model loading logic.
Thanks again for your hard work on this workflow! Any advice or update would be greatly appreciated.

Best regards,

The only working model for me is Qwen image edit 2511