Code for the paper "CREPE: Learnable Prompting With CLIP Improves Visual Relationship Prediction" - View it on GitHub