-LoRA was trained specifically for animated looking style, work good with semi realistic model but animated models are recommended such as hassaku or veridium

-Trained with around 50ish picture

For better result, get between 1- 1.3