Mitsua Diffusion One

Oct 20, 2023

Mitsua Diffusion One is a latent text-to-image diffusion model, which is a successor of Mitsua Diffusion CC0.

This model is trained from scratch using only public domain/CC0 or copyright images with permission for use, with using a fixed pretrained text encoder (OpenCLIP ViT-H/14, MIT License).

Interesting project to use only actually licensed images to train a text to image model

↑ up