LLaVE: Large Language and Vision Embedding Models with Hardness-Weighted Contrastive Learning.

Zhibin Lan, Liqiang Niu, Fandong Meng, Jie Zhou, Jinsong Su

Browse the full EMNLP paper archive.