This is an advanced video so I expect you to have a reasonable amount of experience with the ComfyUI interface. It combines techniques of earlier videos into a single workflow. Workflow is below. drive.google.com/file/d/1KQI8...
Quite an incredible workflow you've got there. I love the fact that it has a good amount of manual work involved which really helps overcome the vagaries of inpainting. While it's easy to see how a good many people might think it's just too many steps for inpainting, as a fellow Photoshopper the manual control is a delight. Thank you for sharing this approach. It's refreshingly different.
Thanks! Before now I have flipped back and forth the P'shop it's only recently that it has become quicker in Comfy. If only the brush in mask editor wasn't so terrible!
I did but not enough to give a final opinion, it seemed a little better than the standard method but it still had problems with the context of the generation. So changing a human head into a robot was fine as they were both heads, but changing a head into a cabbage was more tricky.
Yes, I was a Photoshop professional for thirty years, I often make the elements of an image and then composite, however the round trips to adjust and regenerate parts if you want an adjustment is slower. Seeing them in place immediately makes picture making smoother. With this flow all the parts and variations are available for fine tuning in photoshop when you are finishing off.
@@robadams2451 I meant, you don't need to make a round trips with Auto-Photoshop-StableDiffusion-Plugin. It allows to make all compositing and corrections in the photoshop. I only switch back to comfyui for upscaling as it offers superior features like Mixture of Diffusers.