Uses the Waifu Diffusion model as a base, linked here: https://huggingface.co/hakurei/waifu-diffusion Custom Dreambooth model based off of the likeness of Holo from Spice and Wolf. Dataset was 7 training images, and 14 regularization images. Trained for 3000 steps.
To use the model, simply insert the name "Holo" into your prompts. The class token was "wolf_girl". Add the class token after "Holo" from stronger results. EXAMPLE: "A photo of Holo wolf_girl"