This model was trained on the visual and aesthetic sensibility of the movie "Poor Things." A lot of the objects in the scene are molded around that interpretation, which is something you just can't do with traditional text to image.

→ View original post on X — @skirano