[Illustrious XL] Niji | Retro Anime Style LoRA

LoRA trained with over 200 public images generated by the model.

Recommneded Parameters:

•Strength: 0.9-1.0

•Sampling Method: Euler a

•CFG: 5.0-7.0

•Steps: 25-30

All images shown were upscaled with 4x-UltraSharp.