A textual inversion trained in the style of Ume Aoki, best known as the character designer for Madoka Magica, using the NAI model.
Recommend adding 'border, text' to the negatives (unless you want these things).
Prefer LoRA?
https://civarchive.com/models/110402/ume-aoki-lora
You can also find me on pixai.
Description
Large number of training steps were needed to better emulate the style.