Trained on 21 pics (1280x1280 Color artworks)
Style LoRA based on UDON-YA pics, not trained on doujinshi pictures only colored artworks.