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

Open in new window