arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2208.01944cs.CV
收藏

PalQuant:在低精度加速器上加速高精度网络

PalQuant: Accelerating High-precision Networks on Low-precision Accelerators

Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);Shanghai Jiao Tong University(上海交通大学);Baidu Inc.(百度公司)

Qinghao Hu, Gang Li, Qiman Wu, Jian Cheng

摘要:

近年来,低精度深度学习加速器(DLA)由于其在芯片面积和能耗方面的优势而变得流行,但这些DLA上的低精度量化模型带来了严重的精度下降。实现高精度和高效推理的一种方法是将高精度神经网络部署在低精度DLA上,这很少被研究。在本文中,我们提出了一种低精度量化(PalQuant)方法,该方法通过从头开始学习并行低精度表示来近似高精度计算。此外,我们提出了一种新的循环洗牌模块,以提高并行低精度组之间的跨组信息通信。大量的实验表明,PalQuant在准确性和推理速度方面都优于最先进的量化方法,例如,对于ResNet-18网络量化,PalQuant可以在最先进的2位加速器上同时获得比其4位对应部分高0.52\%的精度和1.78 $\times$加速比。代码可在\url{https://github.com/huqinghao/PalQuant}上获得。

英文摘要:

Recently low-precision deep learning accelerators (DLAs) have become popular due to their advantages in chip area and energy consumption, yet the low-precision quantized models on these DLAs bring in severe accuracy degradation. One way to achieve both high accuracy and efficient inference is to deploy high-precision neural networks on low-precision DLAs, which is rarely studied. In this paper, we propose the PArallel Low-precision Quantization (PalQuant) method that approximates high-precision computations via learning parallel low-precision representations from scratch. In addition, we present a novel cyclic shuffle module to boost the cross-group information communication between parallel low-precision groups. Extensive experiments demonstrate that PalQuant has superior performance to state-of-the-art quantization methods in both accuracy and inference speed, e.g., for ResNet-18 network quantization, PalQuant can obtain 0.52\% higher accuracy and 1.78$\times$ speedup simultaneously over their 4-bit counter-part on a state-of-the-art 2-bit accelerator. Code is available at \url{https://github.com/huqinghao/PalQuant}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑