Optimization design of binary VGG convolutional neural network accelerator
Most of the existing researches on accelerators of binary convolutional neural networks based on FPGA are aimed at small-scale image input, while the applications mainly take large-scale convolutional neural networks such as YOLO and VGG as backbone networks. The hardware of convolutional neural net...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | zho |
Published: |
National Computer System Engineering Research Institute of China
2021-02-01
|
Series: | Dianzi Jishu Yingyong |
Subjects: | |
Online Access: | http://www.chinaaet.com/article/3000128918 |
id |
doaj-765696937da34157b9a0d203378aa903 |
---|---|
record_format |
Article |
spelling |
doaj-765696937da34157b9a0d203378aa9032021-05-21T05:34:34ZzhoNational Computer System Engineering Research Institute of ChinaDianzi Jishu Yingyong0258-79982021-02-01472202310.16157/j.issn.0258-7998.2012073000128918Optimization design of binary VGG convolutional neural network acceleratorZhang Xuxin0Zhang Jia1Li Xinzeng2Jin Jie3College of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201600,ChinaCollege of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201600,ChinaCollege of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201600,ChinaCollege of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201600,ChinaMost of the existing researches on accelerators of binary convolutional neural networks based on FPGA are aimed at small-scale image input, while the applications mainly take large-scale convolutional neural networks such as YOLO and VGG as backbone networks. The hardware of convolutional neural network is optimized and designed from the two aspects including the network topology and pipeline stage, so as to solve the bottleneck of logic resources and improve the performance of the binary VGG network accelerator. CIFAR-10 dataset resized to 224×224 was used to verify the optimized design of VGG convolutional neural network accelerator based on FPGA. Experimental results showed that the system achieved 81% recognition accuracy and 219.9 FPS recognition speed,which verified the effectiveness of the optimization method.http://www.chinaaet.com/article/3000128918optimization designbinary convolutional neural networkfpga accelerator |
collection |
DOAJ |
language |
zho |
format |
Article |
sources |
DOAJ |
author |
Zhang Xuxin Zhang Jia Li Xinzeng Jin Jie |
spellingShingle |
Zhang Xuxin Zhang Jia Li Xinzeng Jin Jie Optimization design of binary VGG convolutional neural network accelerator Dianzi Jishu Yingyong optimization design binary convolutional neural network fpga accelerator |
author_facet |
Zhang Xuxin Zhang Jia Li Xinzeng Jin Jie |
author_sort |
Zhang Xuxin |
title |
Optimization design of binary VGG convolutional neural network accelerator |
title_short |
Optimization design of binary VGG convolutional neural network accelerator |
title_full |
Optimization design of binary VGG convolutional neural network accelerator |
title_fullStr |
Optimization design of binary VGG convolutional neural network accelerator |
title_full_unstemmed |
Optimization design of binary VGG convolutional neural network accelerator |
title_sort |
optimization design of binary vgg convolutional neural network accelerator |
publisher |
National Computer System Engineering Research Institute of China |
series |
Dianzi Jishu Yingyong |
issn |
0258-7998 |
publishDate |
2021-02-01 |
description |
Most of the existing researches on accelerators of binary convolutional neural networks based on FPGA are aimed at small-scale image input, while the applications mainly take large-scale convolutional neural networks such as YOLO and VGG as backbone networks. The hardware of convolutional neural network is optimized and designed from the two aspects including the network topology and pipeline stage, so as to solve the bottleneck of logic resources and improve the performance of the binary VGG network accelerator. CIFAR-10 dataset resized to 224×224 was used to verify the optimized design of VGG convolutional neural network accelerator based on FPGA. Experimental results showed that the system achieved 81% recognition accuracy and 219.9 FPS recognition speed,which verified the effectiveness of the optimization method. |
topic |
optimization design binary convolutional neural network fpga accelerator |
url |
http://www.chinaaet.com/article/3000128918 |
work_keys_str_mv |
AT zhangxuxin optimizationdesignofbinaryvggconvolutionalneuralnetworkaccelerator AT zhangjia optimizationdesignofbinaryvggconvolutionalneuralnetworkaccelerator AT lixinzeng optimizationdesignofbinaryvggconvolutionalneuralnetworkaccelerator AT jinjie optimizationdesignofbinaryvggconvolutionalneuralnetworkaccelerator |
_version_ |
1721432441152864256 |