NACL: A General and Effective KV Cache Eviction Framework for LLMs at Inference Time
This addresses the cost-prohibitive hosting of LLMs for applications requiring extended context windows, representing a strong specific gain in efficiency.
The paper tackles the high memory consumption of KV Cache in large language models during long-context inference by proposing NACL, a framework for evicting unnecessary tokens, which improves performance on short- and long-text tasks by 80% and 76% respectively while reducing KV Cache by up to 50% with over 95% performance maintenance.
Large Language Models (LLMs) have ignited an innovative surge of AI applications, marking a new era of exciting possibilities equipped with extended context windows. However, hosting these models is cost-prohibitive mainly due to the extensive memory consumption of KV Cache involving long-context modeling. Despite several works proposing to evict unnecessary tokens from the KV Cache, most of them rely on the biased local statistics of accumulated attention scores and report performance using unconvincing metric like perplexity on inadequate short-text evaluation. In this paper, we propose NACL, a general framework for long-context KV cache eviction that achieves more optimal and efficient eviction in a single operation during the encoding phase. Due to NACL's efficiency, we combine more accurate attention score statistics in PROXY TOKENS EVICTION with the diversified random eviction strategy of RANDOM EVICTION, aiming to alleviate the issue of attention bias and enhance the robustness in maintaining pivotal tokens for long-context modeling tasks. Notably, our method significantly improves the performance on short- and long-text tasks by 80% and 76% respectively, reducing KV Cache by up to 50% with over 95% performance maintenance. The code is available at https://github.com/PaddlePaddle/Research/tree/master/NLP/ACL2024-NACL.