About

The style LoRA of artist Misasagi Yuuya.

Hope you like it. Follow for more! Have fun!

Usage

• v1:
  style trigger: @misasagi_yuuya
  optional: chibi

• v1.2:

• trigger: @misasagi_yuuya

• optional positive/negative: chibi, oekaki, fewer digits

// "oekaki" helps to generate this high-res pixel art.
// So I do not recommend using "pixel art" tag.

Versions

• v1: Trained on 15 images (9 original and 6 regular). Contains old regularization dataset so I do not recommend using this LoRA with other my styles and characters older than this one.
  If you like this style I will eventually retrain the model with better regular images.

• v1.2: Retrained with 9 synthetic regular images for better generation control and less problems. It is better at replicating author's style.