Generate image from text by guiding a denoising diffusion model. Inference is somewhat slow.

The predecessor to DALLE-2, GLIDE (filtered) with faster PRK/PLMS sampling.

Use stable diffusion and aesthetic CLIP embeddings to guide boring outputs to be more aesthetically pleasing.

Generate 768px images from text using CompVis `retrieval-augmented-diffusion`

CompVis `latent-diffusion text2im` finetuned for inpainting.

Generate speech from text, clone voices from mp3 files. From James Betker AKA "neonbjb".

GLIDE-text2im w/ humans and experimental style prompts.

Guide a StyleGAN3 trained on pictures of mannequins with CLIP.