FLUX.1 INPAINTING
V3.0:
Original：
input2: Use the segmentanything model to segment the mask according to the prompt (keep the details);
input3: Use the segmentanything model to segment the mask according to the prompt (completely replace);
V2.0:
Added the ability to use the character Lora to fix images
Description
FAQ
Comments (5)
I keep getting:
easy imageToMask
Cannot handle this data type: (1, 1, 1152, 4), |u1
What seems to be the issue? it's pointing to the node where I need to select red, green and blue
100%|██████████| 30/30 [01:18<00:00, 2.62s/it]
Warning torch.load doesn't support weights_only on this pytorch version, loading unsafely.
Processing image with shape: 1496 x 1992 to 2992 x 3984
Starting restore_face with codeformer_fidelity: 0.5
Failed inference for CodeFormer: ImageModelDescriptor.__call__() got an unexpected keyword argument 'w'
Failed inference for CodeFormer: ImageModelDescriptor.__call__() got an unexpected keyword argument 'w'
Can you explain the workflow? I have tried a lot of settings but I dont know how to use this. You should explain more details like how to use it or which setting affects what. Most people are no Comfy experts. Dunno how to use the SAM element.
Related documents about this node: https://github.com/chflame163/ComfyUI_LayerStyle/tree/main#segmentanythingultra
The workflow contains nodes that are mostly not connected, can you share a picture of the workflow as is in ComfyUI? Many thanks