A Survey of Low-bit Large Language Models: Basics, Systems, and Algorithms
It addresses the practical deployment challenges of LLMs for researchers and practitioners by summarizing existing methods, but it is incremental as it surveys rather than introduces new techniques.
This paper provides a comprehensive survey of low-bit quantization methods for large language models (LLMs), covering fundamental principles, system implementations, and algorithmic strategies to reduce memory and computational demands, offering insights for enhancing LLM efficiency and applicability.
Large language models (LLMs) have achieved remarkable advancements in natural language processing, showcasing exceptional performance across various tasks. However, the expensive memory and computational requirements present significant challenges for their practical deployment. Low-bit quantization has emerged as a critical approach to mitigate these challenges by reducing the bit-width of model parameters, activations, and gradients, thus decreasing memory usage and computational demands. This paper presents a comprehensive survey of low-bit quantization methods tailored for LLMs, covering the fundamental principles, system implementations, and algorithmic strategies. An overview of basic concepts and new data formats specific to low-bit LLMs is first introduced, followed by a review of frameworks and systems that facilitate low-bit LLMs across various hardware platforms. Then, we categorize and analyze techniques and toolkits for efficient low-bit training and inference of LLMs. Finally, we conclude with a discussion of future trends and potential advancements of low-bit LLMs. Our systematic overview from basic, system, and algorithm perspectives can offer valuable insights and guidelines for future works to enhance the efficiency and applicability of LLMs through low-bit quantization.