The increasing size of large language models has posed challenges for deployment and raised concerns about environmental impact due to high energy consumption. In this work, we introduce BitNet, a scalable and stable 1-bit Transformer architecture designed for large language models. Specifically, we introduce BitLinear as a drop-in replacement of the nn.Linear layer in order to train 1-bit weights from scratch. Experimental results on language modeling show that BitNet achieves competitive performance while substantially reducing memory footprint and energy consumption, compared to state-of-the-art 8-bit quantization methods and FP16 Transformer baselines. Furthermore, BitNet exhibits a scaling law akin to full-precision Transformers, suggesting its potential for effective scaling to even larger language models while maintaining efficiency and performance benefits.
@article{arxiv.2310.11453,
title = {BitNet: Scaling 1-bit Transformers for Large Language Models},
author = {Hongyu Wang and Shuming Ma and Li Dong and Shaohan Huang and Huaijie Wang and Lingxiao Ma and Fan Yang and Ruiping Wang and Yi Wu and Furu Wei},
journal= {arXiv preprint arXiv:2310.11453},
year = {2023}
}