1-bit AI Infra: Part 1.1, Fast and Lossless BitNet b1.58 Inference on CPUs

View a PDF of the paper titled 1-bit AI Infra: Part 1.1, Fast and Lossless BitNet b1.58 Inference on CPUs, by Jinheng Wang and 7 other authors

View PDF
HTML (experimental)

Abstract:Recent advances in 1-bit Large Language Models (LLMs), such as BitNet and BitNet b1.58, present a promising approach to enhancing the efficiency of LLMs in terms of speed and energy consumption. These developments also enable local LLM deployment across a broad range of devices. In this work, we introduce this http URL, a tailored software stack designed to unlock the full potential of 1-bit LLMs. Specifically, we develop a set of kernels to support fast and lossless inference of ternary BitNet b1.58 LLMs on CPUs. Extensive experiments demonstrate that this http URL achieves significant speedups, ranging from 2.37x to 6.17x on x86 CPUs and from 1.37x to 5.07x on ARM CPUs, across various model sizes. The code is available at this https URL.