Open wodsoe opened 3 months ago
it's VERY hard. you can't get it 100% but i've found if you leverage smzSettings along with Auto1111 cliptextencode nodes... you can get 90% identical. also depends the model. with Pony Diffusion i'm developing a workflow that returns 99% identical to Auto1111... however in other models it's around 85%.
note. this doesn't mean quality. I mean just same image. same seed. same image. same pose and composure.
quality is based on other stuff and out the box comfyui produces equal quality to Automatic1111
How to generate same image as a1111 webui? Is there any example workflow?