This Lora is intendet to get better and more realistic Fat behavior
works best with values between 1-1.6
how to get best images:
workflow:
1. Gen image in txt2img
2. regen image in img2img with same prompts, denoise strength ~ 0.60, scale by 1.5-2
Description
FAQ
Comments (2)
So is it not possible for this to work during normal image generation, does it only function in post? I've never really seen that restriction before.
it does work in normal image generation. the img2img just makes it more detailed