CoVLM: Composing Visual Entities and Relationships in Large Language Models Via Communicative Decoding

Published in ICLR, 2024