Just a simple LoRA, trained on om (nk2007)'s art, using the NAI model. It's maybe a little overcooked at 40 epochs, but this proved to be the most stable checkpoint, and it confers his style rather well.
