My second character lora is Ahri's default skin from League of Legends. I had another one previously ready, but it didn't turn out as I wanted so I decided just train it again, using NAI this time, since I already had the dataset ready.
Tail color is a bit iffy, and will sometimes just randomly change.
I also decided to not use regularization images for now, since idk how to use them and they make my lora worse.
Use weight 0.7-1
Use ADetailer extension if you have problems with the face.
I didn't really do much testing, so just try it out.
Feel free to upload your results here.
Any feedback is appreciated.
This lora was trained with 79 images using NAI model.