decided to actually make stable diffusion

r/StableDiffusion
Generative AI AI Research

Creates 48x48 images, with a bidirectional tranformer encoder, trained on flickr8k (and some imagenet), its early in