Distinctive Image Captioning: Leveraging Ground Truth Captions in CLIP Guided Reinforcement Learning