LogQuant: Log-Distributed 2-Bit Quantization of KV Cache with Superior Accuracy Preservation

CHEN Han, Zicong Jiang, Zining Zhang, Bingsheng He, Luo Pingyi, Mian Lu, Yuqiang Chen

Track: long paper (up to 9 pages)

Keywords: Large Language Models, KV Cache, Memory Efficiency, Quantization

Abstract:

We introduce LogQuant, a groundbreaking 2-bit quantization technique for KV Cache in large language model (LLM) inference, delivering substantial memory savings while preserving superior performance. Previous methods either assume that later tokens are more important or attempt to predict important tokens based on earlier attention patterns. Both approaches, however, can result in performance bottlenecks or frequent mispredictions.

LogQuant takes a different approach. By applying a log-based filtering mechanism, it selectively compresses the KV Cache across the entire context, achieving better performance with the same or even reduced memory footprint compared to existing methods. In benchmark tests, it enhances throughput by 25% and boosts batch size by 60% without increasing memory consumption. For challenging tasks such as Math and Code Completion, LogQuant improves accuracy by 40% to 200% at the same compression ratio, outperforming comparable techniques. LogQuant integrates effortlessly with popular inference frameworks like Python’s transformers library and will be made open-source upon publication.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Presenter: ~CHEN_Han2

Format: Yes, the presenting author will attend in person if this work is accepted to the workshop.

Funding: No, the presenting author of this submission does *not* fall under ICLR’s funding aims, or has sufficient alternate funding.

Submission Number: 15