StableMix

Models that are "as realistic as possible without going beyond the boundaries of illustration.

This was created by merging a number of models, including (but not limited to) Waifu diffusion, ChiloutMix, CounterFelt, Nabylon, QuinceMix, and others based on Anything V3.

We are conscious of keeping the shape of the hands and hair stable. It also allows us to better distinguish the boundaries between things!

This will jump-start the success rate of image generation if the appropriate input is given.

This model can handle both general illustrations like AnythingV3 and photo-realistic illustrations.

VAE is built in
Language(s): English

License: The CreativeML OpenRAIL M license

Below is my work generated using this model.