The results of my first experiments with a Google Colab LoRA maker. Results will vary, but I managed to get some improvement over the iterations.
Anime tags were used in the LoRA making process, I suggest anime-style and cartoon-style models, though semi-realistic ones seem to work okay as well.
Other tags to use in prompts:
red nose, demon, horns, claws, black fur, yellow sclera, red eyes, fewer digits, demon tail
Tags for use in the negative prompt:
tongue out, missing horns, missing tail, extra horns, extra tail
