CoVLM: Composing Visual Entities and Relationships in Large Language Models Via Communicative Decoding Paper • 2311.03354 • Published Nov 6, 2023 • 4