DB-LLM: Accurate Dual-Binarization for Efficient LLMs
Preprint
Abstract
Large language models (LLMs) have significantly advanced the field of natural language processing, while the expensive memory and computation consumption impede their practical deployment. Quantization emerges as one of the most effective methods for improving the computational efficiency of LLMs. However, existing ultra-low-bit quantization always causes severe accuracy drops. In this paper, we empirically relieve the micro and macro characteristics of ultra-low bit quantization and present a novel Dual-Binarization method for LLMs, namely DB-LLM. For the micro-level, we take both the accuracy advantage of 2-bit-width and the efficiency advantage of binarization into account, introducing Flexible Dual Binarization (FDB). By splitting 2-bit quantized weights into two independent sets of binaries, FDB ensures the accuracy of representations and introduces flexibility, utilizing the efficient bitwise operations of binarization while retaining the inherent high sparsity of ultra-low bit quantization. For the macro-level, we find the distortion that exists in the prediction of LLM after quantization, which is specified as the deviations related to the ambiguity of samples. We propose the Deviation-Aware Distillation (DAD) method, enabling the model to focus differently on various samples. Comprehensive experiments show that our DB-LLM not only significantly surpasses the current State-of-The-Art (SoTA) in ultra-low bit quantization (eg, perplexity decreased from 9.64 to 7.23), but also achieves an additional 20\% reduction in computational consumption compared to the SOTA method under the same bit-width. Our code will be released soon.Keywords
All Related Versions
This publication has 14 references indexed in Scilit:
- AWQ: Activation-aware Weight Quantization for On-Device LLM Compression and AccelerationGetMobile: Mobile Computing and Communications, 2025
- Revisiting Knowledge Distillation for Autoregressive Language Models2024
- Zero-shot Sharpness-Aware Quantization for Pre-trained Language Models2023
- WinoGrandeCommunications of the ACM, 2021
- ReActNet: Towards Precise Binary Neural Network with Generalized Activation Functions2020
- PIQA: Reasoning about Physical Commonsense in Natural LanguageProceedings of the AAAI Conference on Artificial Intelligence, 2020
- TernaryBERT: Distillation-aware Ultra-low Bit BERT2020
- HellaSwag: Can a Machine Really Finish Your Sentence?2019
- Bi-Real Net: Enhancing the Performance of 1-Bit CNNs with Improved Representational Capability and Advanced Training Algorithm2018
- XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks2016