Relax Within, Balance Across: Geometry-Guided Load Balancing for Vision-Language Mixture-of-Experts Paper • 2608.00574 • Published 4 days ago • 6
Self-Improving Language Models with Bidirectional Evolutionary Search Paper • 2605.28814 • Published May 27 • 62
LOOK-M: Look-Once Optimization in KV Cache for Efficient Multimodal Long-Context Inference Paper • 2406.18139 • Published Jun 26, 2024 • 2