//=time() ?>
The idea was to avoid how anime AI have some sort of pattern of how Diffusion model looks.Make style about ten years back, having animation making anime more 2020s.
UPDATE!Made a mistake it was StyleGAN 2(non ada) not StyleGAN 1. Share the outputs of StyleGAN 2 that feed into IMG2IMG strength 0.5.
火屬性
#stablediffusion Another StyleGAN to StableDiffusion IMG2IMG test. Works very good with vector style.
Others
Soft moe RPG
Heavy painting RPG
Battleship concepts
Boy characters
2020/12 several month I see how far I can get for Tachi-e(立繪) Now try mix Img2Img https://t.co/XrH9A9f7k6