Text-to-image diffusion models use text input to create images. The produced images begin as visual noise that’s gradually refined, with the goal of resembling the user’s prompt. Text input is used to set the scene, specify a subject and can be used to target a particular art style. Here are some examples I generated.
