Download
1 variant available
A LORA which outputs a more anime or concept art look. The model was trained on 350 anime and artstation images. It was trained on the base SDXL 1.0 model for 1 epoch, and 12 batches, resulting in the ~12,248 steps being batched into 1020 steps.
The keyword is [anime]
I would recommend using an anime checkpoint such as animechangefulXL with this lora. Setting the lora around 0.5 keeps the prompt, set it to 2 and it will change the style and possibly the prompt wildly.