my standard settings are typically:
Guidance Scale: 4.1
LoRA Strength: 0.6
Inference Steps: 30
Sampler: euler a
Model: mightymouseV2
Pony:
Sampler: Euler A
Model: PonyRealism
Guidance Scale: 4.1
Steps: 60
Clip Skip: 2
Lora strength: 7-8
i am using easy diffusion if that matters
Description
FAQ
Comments (1)
you should make loras that work at 1.0, or they don't play wel lwith regional prompter, backgrounds, other loras. It's just a big overfit. They become poor, not very responsive. That's why you have to use euler, because it's not very responsive and Euler's a bit more flexible...