May 23, 2022
Researchers present Imagen, combining a large text encoder with diffusion models to generate images from descriptions.
Language understanding and image synthesis develop together. This research publication is separate from later public Imagen products.