The provided texts discuss LMCache, an open-source library designed to enhance the efficiency of large language models (LLMs) by optimizing Key-Value (KV) cache management. A significant innovation highlighted is CacheBlend, a technique integrated into LMCache that drastically improves KV cache hit rates in retrieval-augmented generation (RAG) applications by enabling the reuse of non-prefix texts. This leads to substantial reductions in time to first token (TTFT) and increased throughput, while maintaining high generation quality. The documentation further details LMCache's capabilities, including KV cache offloading to various storage types, sharing across LLMs, and its deployment in production environments like Kubernetes.

Sources:

1) March 1, 2025 - https://blog.lmcache.ai/2025-03-31-eurosys/ - CacheBlend (Best Paper @ ACM EuroSys'25): Enabling 100% KV Cache Hit Rate in RAG

2) https://docs.lmcache.ai/