Layer-Condensed KV Cache for Efficient Inference of Large Language Models

Wu, Haoyi, Tu, Kewei · arXiv · 2024