Benutzer-LoRA
Trained on images from the Tsubaki contest with multiple characters to try to make better characters on a model that supports LORAs.
verdantcross#0
MultiChar