Select a model version, then generate images from text or upload an image to edit it based on instructions.