NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache

Donghyun Son, Euntae Choi, Sungjoo Yoo

Keywords: KV cache quantization, vector quantization, large language model, inference optimization

Abstract:

Large Language Model (LLM) inference is typically memory-intensive, especially when processing large batch sizes and long sequences, due to the large size of key-value (KV) cache.
Vector Quantization (VQ) is recently adopted to alleviate this issue, but we find that the existing approach is susceptible to distribution shift due to its reliance on calibration datasets.
To address this limitation, we introduce $\textbf{NSNQuant}$, a calibration-free Vector Quantization (VQ) technique designed for low-bit compression of the KV cache.
By applying a three-step transformation—$\textbf{1)}$ a token-wise normalization ($\textbf{N}$ormalize), $\textbf{2)}$ a channel-wise centering ($\textbf{S}$hift), and $\textbf{3)}$ a second token-wise normalization ($\textbf{N}$ormalize)—with Hadamard transform, NSNQuant effectively aligns the token distribution with the standard normal distribution.
This alignment enables robust, calibration-free vector quantization using a single reusable codebook.
Extensive experiments show that NSNQuant consistently outperforms prior methods in both 1-bit and 2-bit settings, offering strong generalization and up to 3$\times$ throughput gains over full-precision baselines.

Primary Area: Deep learning (e.g., architectures, generative models, optimization for deep networks, foundation models, LLMs)

Submission Number: 22806