BTC-LLM: Efficient Sub-1-Bit LLM Quantization via Learnable Transformation and Binary Codebook

TL;DR

BTC-LLM introduces a novel sub-1-bit quantization method for large language models, combining binary pattern clustering and learnable transformations to achieve high compression with minimal performance loss.

Contribution

The paper proposes a new sub-1-bit quantization framework that overcomes performance and hardware limitations of prior methods through binary codebook clustering and learnable weight transformation.

Findings

Achieves state-of-the-art sub-1-bit compression (0.7-1.11 bits) on LLMs.

Maintains high accuracy with only 3.1% drop at 0.8 bits on LLaMA-2-13B.

Delivers 1.6x inference speedup over FP16 models.

Abstract

Binary quantization represents the most extreme form of compression, reducing weights to +/-1 for maximal memory and computational efficiency. While recent sparsity-aware binarization achieves sub-1-bit compression via weight pruning, it faces critical challenges: performance degradation, mask-management overhead, and limited hardware compatibility. In this paper, we present BTC-LLM, a novel sub-1-bit LLM quantization framework that leverages binary pattern clustering and weight transformation to overcome these limitations. Our approach incorporates two key innovations: (1) a Binary Codebook that clusters recurring vectors into compact indices using custom distance metrics and sign-based updates; (2) a Learnable Transformation that reduces outliers and promotes shared sign patterns among binary weights. This eliminates sparse masks, enabling efficient inference on standard hardware. Extensive evaluations across LLaMA, Qwen, and FBI-LLM families demonstrate that BTC-LLM achieves state-of-the-art results in extreme compression (1.11-0.7 bits). Notably, BTC-LLM compressed to 0.8 bits on LLaMA-2-13B maintains high performance, with only a 3.1 percent accuracy drop in zero-shot benchmarks, while delivering a 1.6x speedup over FP16.