[1]周密,张维纬,陶英杰,等.采用可替代滤波器的卷积神经网络模型剪枝方法[J].华侨大学学报(自然科学版),2022,43(2):245-251.[doi:10.11830/ISSN.1000-5013.202011013]
　ZHOU Mi,ZHANG Weiwei,TAO Yingjie,et al.Pruning Method of Convolutional Neural Network Using Replaceable Filter[J].Journal of Huaqiao University(Natural Science),2022,43(2):245-251.[doi:10.11830/ISSN.1000-5013.202011013]
点击复制

采用可替代滤波器的卷积神经网络模型剪枝方法()
分享到：

《华侨大学学报（自然科学版）》[ISSN:1000-5013/CN:35-1079/N]

卷:
第43卷
期数:
2022年第2期
页码:
245-251
栏目:
出版日期:
2022-03-08

文章信息/Info

Title:
Pruning Method of Convolutional Neural Network Using Replaceable Filter
文章编号:
1000-5013(2022)02-0245-07
作者:
周密1; 2;  张维纬 1; 2;  陶英杰1; 2;  余浩然1; 2
1. 华侨大学 工学院, 福建 泉州 362021;2. 华侨大学 工业智能化与系统福建省高校工程研究中心, 福建 泉州 362021
Author(s):
ZHOU Mi1; 2;  ZHANG Weiwei 1; 2;  TAO Yingjie1; 2;  YU Haoran1; 2
1. College of Engineering, Huaqiao University, Quanzhou 362021, China; 2. Industrial Intelligence and System Fujian University Engineering Research Center, Huaqiao University, Quanzhou 362021, China
关键词:
剪枝方法;  神经网络模型;  滤波器;  深度学习;  强化学习;  边缘智能
Keywords:
pruning method;  neural network model;  filter;  deep learning;  reinforcement learning;  edge intelligence
分类号:
TP183;TP391.41
DOI:
10.11830/ISSN.1000-5013.202011013
文献标志码:
A
摘要:
将卷积神经网络模型中某一层的所有滤波器抽象到一个欧几里德空间,对其中能被其他滤波器共同表示的滤波器剪枝,降低滤波器冗余,避免精度损失.使用强化学习进行边训练边剪枝,经过微调恢复神经网络模型性能.结果表明:剪枝并微调后的神经网络模型精度损失较小,参数量与浮点计算量显著减少.
Abstract:
All filters in a certain layer of the convolutional neural network are abstracted into an Euclidean space, Pruning the filter that can be jointly represented by other filters, reducing the redundancy of filter and avoiding the loss of accuracy. Using reinforcement learning to prune while training, the neural network model performance is restored through fine-tuning. The results show that, after pruning and fine-tuning, the loss of neural network model accuracy is smaller, the calculated amount of parameters and floating-point are significantly reduced.

参考文献/References:

[1] ZHOU Zhi,CHEN Xu,LI En,et al.Edge intelligence: Paving the last mile of artificial intelligence with edge computing[J].Proceedings of the IEEE,2019,107(8):1738-1762.DOI:10.1109/JPROC.2019.2918951.
[2] GUO Yiwen,YAO Anbang,CHEN Yurong.Dynamic network surgery for efficient dnns[C]//Advances in Neural Information Processing Systems.Montreal:[s.n.],2016:1379-1387.
[3] CARREIRA-PERPINáN M A,IDELBAYE V Y.“Learning-Compression” algorithms for neural net pruning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake:IEEE Press,2018:8532-8541.
[4] DING Xiaohan,ZHOU Xiangxin,GUO Yuchen,et al.Global sparse momentum SGD for pruning very deep neural networks[C]//Advances in Neural Information Processing Systems.Vancouver:[s.n.],2019:6382-6394.
[5] HE Yihui,ZHANG Xiangyu,SUN Jian.Channel pruning for accelerating very deep neural networks[C]//Proceedings of the IEEE International Conference on Computer Vision.Venice:IEEE Press,2017:1389-1397.
[6] HUANG Qiangui,ZHOU K,YOU Suya,et al.Learning to prune filters in convolutional neural networks[C]//Winter Conference on Applications of Computer Vision.Lake Tahoe:IEEE Press,2018:709-718.
[7] YOU Zhonghui,YAN Kun,YE Jinmian,et al.Gate decorator: Global filter pruning method for accelerating deep convolutional neural networks[C]//Conference on Neural Information Processing Systems.Vancouver:[s.n.],2019:2133-2144.
[8] HE Yang,LIU Ping,WANG Ziwei,et al.Filter pruning via geometric median for deep convolutional neural networks acceleration[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE Press,2019:4340-4349.DOI:10.1109/CVPR.2019.00447.
[9] SILVER D,LEVER G,HEESS N,et al.Deterministic policy gradient algorithms[C]//International Conferenceon Machine Learning.Cambridge:MIT Press,2014:387-395.
[10] HE Yihui,LIN Ji,LIU Zhijian,et al.Amc: Automl for model compression and acceleration on mobile devices[C]//Proceedings of the European Conference on Computer Vision.Munich:Springer,2018:784-800.
[11] FLETCHER P T,VENKATASUBRAMANIAN S,JOSHI S.Robust statistics on Riemannian manifolds via the geometric median[C]//2008 IEEE Conference on Computer Vision and Pattern Recognition.Anchrage:IEEE Press,2008:1-8.DOI:10.1109/CVPR.2008.4587747.
[12] HE Kaiming,ZHANG Xiangyu,REN Shaoqing,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE Press,2016:770-778.DOI:10.1109/CVPR.2016.90.
[13] RUSSAKOVSKY O,DENG Jia,SU Hao,et al.Imagenet large scale visual recognition challenge[J].International Journal of Computer Vision,2015,115(3):211-252.DOI:10.1007/s11263-015-0816-y.
　　

备注/Memo

备注/Memo:
收稿日期: 2020-11-04
通信作者: 张维纬(1982-),男,副教授,博士,主要从事大数据、物联网及边缘智能的研究.E-mail:178483968@qq.com.
基金项目: 国家自然科学基金面上资助项目(61976098); 福建省泉州市科技计划项目(2020C067); 华侨大学研究生科研创新能力培育计划项目(18014084015)
更新日期/Last Update: 2022-03-20