Improving Multi-step RAG with Hypergraph-based Memory for Long-Context Complex Relational Modeling Paper • 2512.23959 • Published 25 days ago • 107
UME-R1 Collection UME-R1 is a framework designed to endow multimodal embedding models with the flexibility to switch between discriminative and generative embeddings • 4 items • Updated Nov 4, 2025 • 8
UME-R1: Exploring Reasoning-Driven Generative Multimodal Embeddings Paper • 2511.00405 • Published Nov 1, 2025 • 6
LLaVE: Large Language and Vision Embedding Models with Hardness-Weighted Contrastive Learning Paper • 2503.04812 • Published Mar 4, 2025 • 15
LLaVE Collection LLaVE is a series of large language and vision embedding models trained on a variety of multimodal embedding datasets • 4 items • Updated Mar 10, 2025 • 8