BitDistiller: Unleashing the Potential of Sub-4-Bit LLMs via Self-Distillation
This work addresses the problem of efficient deployment of LLMs for practitioners by offering a more cost-effective solution with reduced data and training needs, though it is incremental as it builds on existing quantization and distillation techniques.
The paper tackles the deployment challenges of large language models by introducing BitDistiller, a framework that combines quantization-aware training with knowledge distillation to enhance performance at ultra-low precisions (sub-4-bit), achieving significant improvements over existing methods in 3-bit and 2-bit configurations on benchmarks.
The upscaling of Large Language Models (LLMs) has yielded impressive advances in natural language processing, yet it also poses significant deployment challenges. Weight quantization has emerged as a widely embraced solution to reduce memory and computational demands. This paper introduces BitDistiller, a framework that synergizes Quantization-Aware Training (QAT) with Knowledge Distillation (KD) to boost the performance of LLMs at ultra-low precisions (sub-4-bit). Specifically, BitDistiller first incorporates a tailored asymmetric quantization and clipping technique to maximally preserve the fidelity of quantized weights, and then proposes a novel Confidence-Aware Kullback-Leibler Divergence (CAKLD) objective, which is employed in a self-distillation manner to enable faster convergence and superior model performance. Empirical evaluations demonstrate that BitDistiller significantly surpasses existing methods in both 3-bit and 2-bit configurations on general language understanding and complex reasoning benchmarks. Notably, BitDistiller is shown to be more cost-effective, demanding fewer data and training resources. The code is available at https://github.com/DD-DuDa/BitDistiller.