V1 Anima: Trained on preview V2. not sure if i'm gonna continue this, i'd have to edit the images in the dataset to get this up to the quality i want and it doesn't seem like a lot of people want it anyway, but anima trains fast so i figured why not train it for it.

V2 ZiT: Fixed the same face issue. much better prompt adherence though not perfect - attempted a demoness with red skin and it just wasn't working out, but it overall will follow prompts much better than before. more diversity in shapes and number of twists as well.

V1 ZiT: Overcooked