Trained with Kohya's Colab pages (Lora_trainer_XL & Dataset_Maker).
Base: Edanoma
Pixiv Artist with a Unique style.
Works best at strength 1. Preview images are made with my other character LoRAs i made.
Optional Tags: ---
Description
FAQ
Comments (1)
For art more accurate to their coloring:
apply the "Spot Color" tag, and only specify one or two colors (for hair or eyes, potentially)
Works best at higher resolutions (1024x1024 worked best for me), 768x768 worked most of the time but sometimes produced smudgy features.