Anyway, I'm late and out of time for the hype as usual
I just hope it's good enough to please some people. I've compared it to others I've seen on CivitaAI and I think it works well, but that's subjective.
Use it at 0.6 and 0.5, also eye colour, in the examples you can see how it varies, all this is because some of the colour prompts are quite strong and go to the eyes, so if you do not like the eyes do not be lazy and write in a prompt "green eyes"
Description
FAQ
Comments (3)
Hi, I can't seem to replicate your sample images using the generation data provided; is there some sort of issue with this LoRA?
I don't know what to say, the generation itself is very much dependent on xformers different versions of cuda and video card model
woke up for you and checked everything, I seem to have something wrong with the art that was made with cetusMIX, sorry for misleading you, may have mixed up the art with one of the other versions, I have already changed them. I have quite a few projects and a very specific sorting methodology