Dynamic-stride-net: deep convolutional neural network with dynamic stride

Zerui Yang; Yuhui Xu; Wenrui Dai; Hongkai Xiong

doi:10.1117/12.2537799

18 November 2019 Dynamic-stride-net: deep convolutional neural network with dynamic stride

Zerui Yang, Yuhui Xu, Wenrui Dai, Hongkai Xiong

Proceedings Volume 11187, Optoelectronic Imaging and Multimedia Technology VI; 1118707 (2019) https://doi.org/10.1117/12.2537799
Event: SPIE/COS Photonics Asia, 2019, Hangzhou, China

Abstract

It is crucial to reduce the cost of deep convolutional neural networks while preserving their accuracy. Existing methods adaptively prune DNNs in a layer-wise or channel-wise manner based on the input image. In this paper, we develop a novel dynamic network, namely Dynamic-Stride-Net, to improve residual network with layer-wise adaptive strides in the convolution operations. Dynamic-Stride-Net leverages a gating network to adaptively select the strides of convolutional blocks based on the outputs of the previous layer. To optimize the selection of strides, the gating network is trained by reinforcement learning. The floating point operations per second (FLOPS) is significantly reduced by adapting the strides to convolutional layers without loss of accuracy. Dynamic-Stride-Net reduces the computational cost by 35%-50% with equivalent accuracy of the original model on CIFAR-10 and CIFAR-100 datasets. It outperforms the state-of-the-art dynamic networks and static compression methods.

Citation Download Citation

Zerui Yang, Yuhui Xu, Wenrui Dai, and Hongkai Xiong "Dynamic-stride-net: deep convolutional neural network with dynamic stride", Proc. SPIE 11187, Optoelectronic Imaging and Multimedia Technology VI, 1118707 (18 November 2019); https://doi.org/10.1117/12.2537799

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available