首页|AutoQNN:An End-to-End Framework for Automatically Quantizing Neural Networks

AutoQNN:An End-to-End Framework for Automatically Quantizing Neural Networks

扫码查看
原文链接
  • 万方数据
Exploring the expected quantizing scheme with suitable mixed-precision policy is the key to compress deep neural networks(DNNs)in high efficiency and accuracy.This exploration implies heavy workloads for domain experts,and an automatic compression method is needed.However,the huge search space of the automatic method introduces plenty of computing budgets that make the automatic process challenging to be applied in real scenarios.In this paper,we propose an end-to-end framework named AutoQNN,for automatically quantizing different layers utilizing different schemes and bitwidths without any human labor.AutoQNN can seek desirable quantizing schemes and mixed-precision policies for mainstream DNN models efficiently by involving three techniques:quantizing scheme search(QSS),quantiz-ing precision learning(QPL),and quantized architecture generation(QAG).QSS introduces five quantizing schemes and defines three new schemes as a candidate set for scheme search,and then uses the Differentiable Neural Architecture Search(DNAS)algorithm to seek the layer-or model-desired scheme from the set.QPL is the first method to learn mixed-precision policies by reparameterizing the bitwidths of quantizing schemes,to the best of our knowledge.QPL optimizes both classification loss and precision loss of DNNs efficiently and obtains the relatively optimal mixed-precision model within limited model size and memory footprint.QAG is designed to convert arbitrary architectures into corresponding quantized ones without manual intervention,to facilitate end-to-end neural network quantization.We have implemented AutoQNN and integrated it into Keras.Extensive experiments demonstrate that AutoQNN can consistently outperform state-of-the-art quantization.For 2-bit weight and activation of AlexNet and ResNet18,AutoQNN can achieve the accura-cy results of 59.75％and 68.86％,respectively,and obtain accuracy improvements by up to 1.65％and 1.74％,respectively,compared with state-of-the-art methods.Especially,compared with the full-precision AlexNet and ResNet18,the 2-bit models only slightly incur accuracy degradation by 0.26％and 0.76％,respectively,which can fulfill practical application demands.

automatic quantizationmixed precisionquantizing scheme searchquantizing precision learningquan-tized architecture generation

龚成、卢冶、代素蓉、邓倩、杜承昆、李涛

展开 >

College of Software,Nankai University,Tianjin 300350,China

College of Computer Science,Nankai University,Tianjin 300350,China

State Key Laboratory of Processors,Institute of Computing Technology,Chinese Academy of Sciences Beijing 100190,China

China Postdoctoral Science FoundationNational Natural Science Foundation of ChinaNational Natural Science Foundation of ChinaSpecial Funding for Excellent Enterprise Technology Correspondent of TianjinOpen Project Foundation of Information Security Evaluation Center of Civil Aviation,Civil Aviation University of China

2022M721707620021756227224821YDTPJC00380ISECCA-202102

2024

10.1007/s11390-022-1632-9

计算机科学技术学报(英文版)
中国计算机学会

计算机科学技术学报(英文版)

CSTPCD
影响因子：0.432
ISSN：1000-9000
年,卷(期)：2024.39(2)
  • 1