Please support Qwen-Image-2512-Fun-ControlNet-Union and Z-Image-Turbo-Fun-Controlnet-Union-2.1...2601

[thaivb]

Feature Idea

Load fron this https://modelscope.cn/models/PAI/Qwen-Image-2512-Fun-Controlnet-Union

2601*********
https://huggingface.co/alibaba-pai/Z-Image-Turbo-Fun-Controlnet-Union-2.1
https://huggingface.co/alibaba-pai/Qwen-Image-2512-Fun-Controlnet-Union

Existing Solutions

No response

Other

No response

[badraymen]

+1

[selfonchik777]

+1

[ishagalov]

+1

[candid80]

+1

[kishudoshi01]

A week since it has been released and still no comyui support... :(

[raphaeldevil]

(｡•ˇ‸ˇ•｡)

[DmitryKripakov]

+1 I am really looking forward to it. :(

[kishudoshi01]

+1

[petmycat]

EDIT: I realized that my earlier implementation was no good, so I tried a different approach that does not change ComfyUI's original core files, instead now the patch exists as a custom node pack which you can find here: https://github.com/petmycat/ComfyUI-gen2

I got so curious of how that controlnet turns out that I made a temporary patch myself, PLEASE BACKUP YOUR ORIGINAL FILES AND USE AT YOUR OWN RISKS! In short, go to ComfyUI root directory and look for folders comfy and comfy_extras, find these files and replace them with mine. Then if you restart ComfyUI and see that the QwenDiffsynthControlnet node has an inpaint_image input, you're good to go.

And I have to point out that I am not a tech-savvy guy so a lot of the coding work was achieved through the help of GPT5.2 Codex. So yeah, BACKUP YOUR FILES!!!

under ComfyUI/comfy:
model_base.py

sd.py

conds.py

under ComfyUI/comfy_extras:
nodes_model_patch.py

Here is an example snapshot of how it should turn out:

[frzfrzz]

+1

[candid80]

I got so curious of how that controlnet turns out that I made a temporary patch myself, PLEASE BACKUP YOUR ORIGINAL FILES AND USE AT YOUR OWN RISKS! In short, go to ComfyUI root directory and look for folders comfy and comfy_extras, find these files and replace them with mine. Then if you restart ComfyUI and see that the QwenDiffsynthControlnet node has an inpaint_image input, you're good to go.

And I have to point out that I am not a tech-savvy guy so a lot of the coding work was achieved through the help of GPT5.2 Codex. So yeah, BACKUP YOUR FILES!!!

under ComfyUI/comfy: model_base.py

sd.py

conds.py

under ComfyUI/comfy_extras: nodes_model_patch.py

Here is an example snapshot of how it should turn out:

You should consider making this a PR.

[kishudoshi01]

I got so curious of how that controlnet turns out that I made a temporary patch myself, PLEASE BACKUP YOUR ORIGINAL FILES AND USE AT YOUR OWN RISKS! In short, go to ComfyUI root directory and look for folders comfy and comfy_extras, find these files and replace them with mine. Then if you restart ComfyUI and see that the QwenDiffsynthControlnet node has an inpaint_image input, you're good to go.

And I have to point out that I am not a tech-savvy guy so a lot of the coding work was achieved through the help of GPT5.2 Codex. So yeah, BACKUP YOUR FILES!!!

under ComfyUI/comfy: model_base.py

sd.py

conds.py

under ComfyUI/comfy_extras: nodes_model_patch.py

Here is an example snapshot of how it should turn out:

Workflow for this please?