It takes an input named `prompt` which contains the image description and outputs an image.