Kanon's lora model.

This model focuses on the simulation and restoration of TV anime.

There is no trigger words, just need to adjust its weight in the additional net. You may need to write"1girl,orange_hair,puple_eyes" in your prompt to active the lora.

When you try to generate TV anime style pictures, the recommended weight is 0.7-0.9. Otherwise 0.4-0.6 is recommended.

This is my first time training a lora model, obviously it's not perfect. If you have anything to said, feel free to comment below. I may update with a better model in a few days.