This model was trained on anime illustration generated by the Dall-E 3 model.
I have decided to separate the versions of it and now the LoRA for AnimagineXL will be hosted separately
FLUX version: https://civarchive.com/models/1996960/dall-e-3-anime-style-flux
AnimagineXL version: https://civarchive.com/models/619600/dall-e-3-anime-style-animaginexl
Description
Since I enjoy the creativity of the AnimagineXL model, I decided to try making a version of my successful LoRA that recreates the Dall-E 3 anime style tailored for it. The result is more noticeable when you use 'soft glowing background' in the prompt.
FAQ
Comments (5)
this is sick! if you dont mind me asking, how many images was this lora trained on?
36
@DOK oh wow thx, trained many characters but ive never done a style before. any tips?
@BakaBlitz its pretty simple, i just use style preset on this site. I think the main thing is selecting the most successful examples and describing them well. For the descriptions, I use ChatGPT 4
@DOK aah i see. i was looking into the style preset actually, wasnt sure if it was good. thanks for the insight!