Self-Annotated Training for Controllable Image Captioning

Open in new window