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

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