CommVQ: Commutative Vector Quantization for KV Cache Compression

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Junyan, Zhang, Yang, Hassan, Muhammad Yusuf, Chafekar, Talha, Cai, Tianle, Ren, Zhile, Guo, Pengsheng, Karimzadeh, Foroozan, Reed, Colorado, Wang, Chong, Gan, Chuang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912445102030848
author Li, Junyan
Zhang, Yang
Hassan, Muhammad Yusuf
Chafekar, Talha
Cai, Tianle
Ren, Zhile
Guo, Pengsheng
Karimzadeh, Foroozan
Reed, Colorado
Wang, Chong
Gan, Chuang
author_facet Li, Junyan
Zhang, Yang
Hassan, Muhammad Yusuf
Chafekar, Talha
Cai, Tianle
Ren, Zhile
Guo, Pengsheng
Karimzadeh, Foroozan
Reed, Colorado
Wang, Chong
Gan, Chuang
contents Large Language Models (LLMs) are increasingly used in applications requiring long context lengths, but the key-value (KV) cache often becomes a memory bottleneck on GPUs as context grows. To address this, we propose Commutative Vector Quantization (CommVQ) to significantly reduce memory usage for long-context LLM inference. We first introduce additive quantization with a lightweight encoder and codebook to compress the KV cache, which can be decoded via simple matrix multiplication. To further reduce computational costs during decoding, we design the codebook to be commutative with Rotary Position Embedding (RoPE) and train it using an Expectation-Maximization (EM) algorithm. This enables efficient integration of decoding into the self-attention mechanism. Our approach achieves high accuracy with additive quantization and low overhead via the RoPE-commutative codebook. Experiments on long-context benchmarks and GSM8K show that our method reduces FP16 KV cache size by 87.5% with 2-bit quantization, while outperforming state-of-the-art KV cache quantization methods. Notably, it enables 1-bit KV cache quantization with minimal accuracy loss, allowing a LLaMA-3.1 8B model to run with a 128K context length on a single RTX 4090 GPU. The source code is available at: https://github.com/UMass-Embodied-AGI/CommVQ.
format Preprint
id arxiv_https___arxiv_org_abs_2506_18879
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CommVQ: Commutative Vector Quantization for KV Cache Compression
Li, Junyan
Zhang, Yang
Hassan, Muhammad Yusuf
Chafekar, Talha
Cai, Tianle
Ren, Zhile
Guo, Pengsheng
Karimzadeh, Foroozan
Reed, Colorado
Wang, Chong
Gan, Chuang
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) are increasingly used in applications requiring long context lengths, but the key-value (KV) cache often becomes a memory bottleneck on GPUs as context grows. To address this, we propose Commutative Vector Quantization (CommVQ) to significantly reduce memory usage for long-context LLM inference. We first introduce additive quantization with a lightweight encoder and codebook to compress the KV cache, which can be decoded via simple matrix multiplication. To further reduce computational costs during decoding, we design the codebook to be commutative with Rotary Position Embedding (RoPE) and train it using an Expectation-Maximization (EM) algorithm. This enables efficient integration of decoding into the self-attention mechanism. Our approach achieves high accuracy with additive quantization and low overhead via the RoPE-commutative codebook. Experiments on long-context benchmarks and GSM8K show that our method reduces FP16 KV cache size by 87.5% with 2-bit quantization, while outperforming state-of-the-art KV cache quantization methods. Notably, it enables 1-bit KV cache quantization with minimal accuracy loss, allowing a LLaMA-3.1 8B model to run with a 128K context length on a single RTX 4090 GPU. The source code is available at: https://github.com/UMass-Embodied-AGI/CommVQ.
title CommVQ: Commutative Vector Quantization for KV Cache Compression
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2506.18879