I'm going to be real here, training this concept was hell. even after all my best efforts the chains are somewhat inconsistent at times, and required quite a bit of thinkering to work. I blame the dataset being only like 60 good images.
naked characters are very consistent, but the more clothes and jewelry you add it gets more and more confused.
recommended strength is 0.9.
Description
FAQ
Comments (6)
Magical 😍
So hot 🥵, awesome
I love this concept and I wish you could make a V2 version.
I've uploaded 50 pictures hoping they are good enough for you to keep working on it ❤️, cheers.
(also would be lovely even just a V shape version)
I don't think I'm making a v2 any time soon, but anyone is free to rise up to the task if they so desire.
soo nice lora
Amazing！