ICQuant: Index Coding enables Low-bit LLM Quantization

TL;DR

ICQuant introduces an efficient index coding scheme leveraging outlier statistics to significantly improve low-bit quantization of LLM weights, enabling high accuracy with minimal bit overhead.

Contribution

The paper proposes ICQuant, a novel outlier-aware quantization framework that reduces bit overhead and enhances low-bit LLM quantization without fine-tuning.

Findings

ICQuant requires only 0.3 bits per weight for outlier suppression.

ICQuant improves zero-shot accuracy of 2-bit Llama3-70B by up to 150%.

ICQuant achieves comparable performance to fine-tuned quantizers without additional training.

Abstract

The rapid deployment of Large Language Models (LLMs) highlights the need for efficient low-bit post-training quantization (PTQ), due to their high memory costs. A key challenge in weight quantization is the presence of outliers, which inflate quantization ranges and lead to large errors. While a number of outlier suppression techniques have been proposed, they either: fail to effectively shrink the quantization range, or incur (relatively) high bit overhead. In this paper, we present ICQuant, a novel framework that leverages outlier statistics to design an efficient index coding scheme for outlier-aware weight-only quantization. Compared to existing outlier suppression techniques requiring $\approx 1$ bit overhead to halve the quantization range, ICQuant requires only $\approx 0.3$ bits; a significant saving in extreme compression regimes (e.g., 2-3 bits per weight). ICQuant can be used on top of any existing quantizers to eliminate outliers, improving the quantization quality. Using just 2.3 bits per weight and simple scalar quantizers, ICQuant improves the zero-shot accuracy of the 2-bit Llama3-70B model by up to 130% and 150% relative to QTIP and QuIP#; and it achieves comparable performance to the best-known fine-tuned quantizer (PV-tuning) without fine-tuning.