Based on SD1.5 , and works with Anime models but was not intended for that. I will train another one based on Anime models later on.
The Lora works well with strength between 0.5 and 0.7; it was trained on images of 768 by 768 so use that resolution preferably. Also, I recommend using Image to Image and control-net for scaling instead of high res fix but is a matter of choice.
Based on SD1.5 , and works with Anime models but was not intended for that. I will train another one based on Anime models later on.
The Lora works well with strength between 0.5 and 0.7; it was trained on images of 768 by 768 so use that resolution preferably. Also, I recommend using Image to Image and control-net for scaling instead of high res fix but is a matter of choice.