Machine learning models which take an input natural language description and produce an image matching that description, e.g. DALL-E, Stable Diffusion, or Midjourney.