ML//diffusion model//DALL-E

DALL-E is OpenAI's text-to-image model. DALL-E 1 (January 2021) combined a discrete VAE with a transformer; DALL-E 2 (April 2022) moved to diffusion with CLIP. It was the model that proved AI creativity had arrived, or at least outputs that feel creative.

DALL-E is OpenAI's text-to-image model. DALL-E 1 (January 2021) combined a discrete VAE with a transformer; DALL-E 2 (April 2022) moved to diffusion with CLIP. It was the model that proved AI creativity had arrived, or at least outputs that feel creative.