What is this?

This LoRA is intended to patch Illustrious checkpoints to produce better furry images from e621 tags.

It works with some better than others but should generally give you more consistent furry images from a checkpoint that would otherwise have more trouble with it.

Usage tips:

Usage depends on the base model a bit, but I recommend starting with Euler A Karras or a DPM++ based sampler, and prompt using mostly E621 tags (a bit of natural language may be okay)