My first LoRA — trained on around 40 images using the Anima 1.0 base.

Main trigger:

RIMURUTRIGGER

I also always add: light blue hair, long hair, yellow eyes,

Just to make sure hair and eye color is right across generations

The rest of the triggers are optional outfits

Recommended strength:

0.6 is what I used for the preview images, but it should work higher or lower depending on your checkpoint and prompt.

Training notes:

Most preview images were generated at 1240x1024 and upscaled 1.5x. The dataset images were created with a Turbo LoRA, but this LoRA works fine without it.

NSFW:

Seems to work fine for NSFW as well.

If you run into issues, let me know. I have earlier and later epoch versions saved and can upload another version if needed.