Training a diffusion model to generate anime/cartoon images. This project was inspired by the fact that current text-to-image models like Dalle-2 or Imagen don't produce the highest quality anime images. So I decided to train a diffusion model (without CLIP and text part) to generate anime/cartoon images.
This project utilizes the Tagged Anime Illustrations dataset from Kaggle, which is not included in the repository and can be found here: https://www.kaggle.com/datasets/mylesoneill/tagged-anime-illustrations