bai asuka artstyle

Trained on Pony Diffusion V6 XL, with a cleaner dataset and LoCon parameters.

There are no trigger words used in this version, but following the prompting "style" in the gallery showcases will give the best results. In addition:

• Use a LoRA weight scale between 0.9 to 1.0

  • The higher the closer it is to Bai Asuka's style, but will clutter the image with speech bubbles, sfx, etc. This is because I wasn't able to clean the dataset much. (my bad)

• Use detailers such as EasyNegative