通过敏感度分析高效量化模型，兼顾速度与精度。
Sensitivity-Aware Post-Training Quantization for Deep Neural Networks
- 按参数敏感度优先量化，低敏感参数保留补偿误差。
- 速度比基线快20至200倍，平均精度损失低于0.3%。
- 适合边缘计算和实时推理等资源受限场景。
模型量化通过降低神经网络参数精度实现压缩，但常导致精度下降。现有后训练量化（PTQ）方法采用迭代参数更新以在高压缩比下保持精度，计算复杂度高、资源开销大，限制了其在资源受限的边缘计算和实时推理场景的应用。本文提出一种基于参数敏感度分析的高效PTQ方法，优先量化高敏感度参数，利用未量化低敏感度参数补偿量化误差，从而减轻精度损失。此外，通过挖掘参数敏感度的列间聚类特性，引入行并行量化框架与全局共享逆海塞矩阵更新机制，使计算复杂度降低一个数量级。在ResNet-50和YOLOv5s上的实验表明，该方法相比Optimal Brain Quantization基线实现20至200倍的量化加速，平均精度损失低于0.3%，验证了其在效率与精度间的良好平衡。
原文摘要 · Abstract (English)
Model quantization reduces neural network parameter precision to achieve compression, but often compromises accuracy. Existing post-training quantization (PTQ) methods employ iterative parameter updates to preserve accuracy under high compression ratios, incurring significant computational complexity and resource overhead, which limits applicability in resource-constrained edge computing and real-time inference scenarios. This paper proposes an efficient PTQ method guided by parameter sensitivity analysis. The approach prioritizes quantization of high-sensitivity parameters, leveraging unquantized low-sensitivity parameters to compensate for quantization errors, thereby mitigating accuracy degradation. Furthermore, by exploiting column-wise clustering of parameter sensitivity, the method introduces a row-parallel quantization framework with a globally shared inverse Hessian matrix update mechanism, reducing computational complexity by an order of magnitude. Experimental results on ResNet-50 and YOLOv5s demonstrate a 20-200-fold quantization speedup over the Optimal Brain Quantization baseline, with mean accuracy loss below 0.3%, confirming the method's efficacy in balancing efficiency and accuracy.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。