USE V2 IF YOU DON'T CARE ABOUT THE MASSIVE EYE MAKEUP
REALLY BAD on stable diffusion 1.5, use something that's better at generating people. For the example images I used genericfemalemix 2.1, hassanblend 1.5, and SD1.5. you'll know when it's SD,,,, still has issues on other models though
https://civitai.com/models/4190/genericfemalemixgefemi
^the latest version is the inpainting model and you prob don't want to use that
the bad images are from SD, the good ones are from genericfemalemix 2.1, the ones that are in cars(?) or have natural colored hair are from hassanblend 1.5, and the anime images are from anything v3
sd1.5 generates significantly larger areas of eye makeup though, which is a lot of fun, but pretty consistently generates really malformed faces so not worth
made this because I wanted it :)))))))))))
USE V2 IF YOU DON'T CARE ABOUT THE MASSIVE EYE MAKEUP
even worse than the old one on stable diffusion, so definitely don't use it there. idk what models to train this on tbh.
trained for like 500 steps on generic female mix, and 500 on sd1.5
Go ahead and upload yours!
Your query returned no results β please try removing some filters or trying a different term.