Compensate Quantization Errors+: Quantized Models Are Inquisitive Learners

TL;DR

This paper introduces novel quantization techniques for large language models that improve performance by refining weight distributions and distributing errors more evenly, outperforming existing methods.

Contribution

The paper proposes Singular-value Diagonal Expansion and Cross-layer Learning to enhance weight quantization, addressing distribution shifts and error distribution issues in LLMs.

Findings

Significant performance gains over state-of-the-art quantization methods.

Effective weight distribution refinement improves quantization accuracy.

Error distribution across layers enhances overall model performance.

Abstract

The quantization of large language models (LLMs) has been a prominent research area aimed at enabling their lightweight deployment in practice. Existing research about LLM's quantization has mainly explored the interplay between weights and activations, or employing auxiliary components while neglecting the necessity of adjusting weights during quantization. Consequently, original weight distributions frequently fail to yield desired results after round-to-nearest (RTN) quantization. Even though incorporating techniques such as mixed precision and low-rank error approximation in LLM's quantization can yield improved results, they inevitably introduce additional computational overhead. On the other hand, traditional techniques for weight quantization, such as Generative Post-Training Quantization, rely on manually tweaking weight distributions to minimize local errors, but they fall short of achieving globally optimal outcomes. Although the recently proposed Learnable Singular-value Increment improves global weight quantization by modifying weight distributions, it disrupts the original distribution considerably. This introduces pronounced bias toward the training data and can degrade downstream task performance. In this paper, we introduce Singular-value Diagonal Expansion, a more nuanced approach to refining weight distributions to achieve better quantization alignment. Furthermore, we introduce Cross-layer Learning that improves overall quantization outcomes by distributing errors more evenly across layers. Our plug-and-play weight-quantization methods demonstrate substantial performance improvements over state-of-the-art approaches, including OmniQuant, DuQuant, and PrefixQuant.