You’d have to combine it with one of the text to image models.