The first LoRA I have ever made.
This LoRA is about a fictional person who does not exist in real life!
All pictures were taken with ComfyUI therefore there is no generation data!
I achieved the best results with around 0.8-0.9.
The result of the clothes depends on how you describe it, since I have trained all the clothes she wore in the anime, you may or may not get fused clothes, pay attention to how you specify the clothes.
I went through the entire first season in full HD and took a total of 154 screenshots.
The images were trained with 768 x 768. The original image was not cropped to this size, I simply took the images as I had screenshoted them.