Extract data from documents into JSON
Apply the motion of a video on a portrait
Train custom Stable Diffusion models