This LoRA tries to evoke Incase's artstyle in Chroma.

I wanted to try out how the onsite training for Chroma works, because I haven't managed to get it running on my system yet - neither for generation nor for training. Crying in 16 MB VRam here... and one of the universally liked artist styles seemed like a good starting point.

The trigger should be beginning a prompt with something like "An image by Incase 2019". The 2019 wasn't originally planned to be a part of it, but got tagged in on the whole dataset...

This dataset is uncensored. I dunno how to get rid of that annoying signature, though... probably have to edit the dataset to be safe.