Accelerating Deep Neural Networks by Combining Block-Circulant Matrices and Low-Precision Weights
School of Electronic Science and Engineering, Nanjing University, Nanjing 210023, China
State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing 210023, China
School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, 114 28 Stockholm, Sweden
Author to whom correspondence should be addressed.
Received: 23 November 2018 / Revised: 16 December 2018 / Accepted: 5 January 2019 / Published: 10 January 2019
As a key ingredient of deep neural networks (DNNs), fully-connected (FC) layers are widely used in various artificial intelligence applications. However, there are many parameters in FC layers, so the efficient process of FC layers is restricted by memory bandwidth. In this paper, we propose a compression approach combining block-circulant matrix-based weight representation and power-of-two quantization. Applying block-circulant matrices in FC layers can reduce the storage complexity from
. By quantizing the weights into integer powers of two, the multiplications in the reference can be replaced by shift and add operations. The memory usages of models for MNIST, CIFAR-10 and ImageNet can be compressed by
with minimal accuracy loss, respectively. A configurable parallel hardware architecture is then proposed for processing the compressed FC layers efficiently. Without multipliers, a block matrix-vector multiplication module (B-MV) is used as the computing kernel. The architecture is flexible to support FC layers of various compression ratios with small footprint. Simultaneously, the memory access can be significantly reduced by using the configurable architecture. Measurement results show that the accelerator has a processing power of 409.6 GOPS, and achieves 5.3 TOPS/W energy efficiency at 800 MHz.
This is an open access article distributed under the Creative Commons Attribution License
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).
Share & Cite This Article
MDPI and ACS Style
Qin, Z.; Zhu, D.; Zhu, X.; Chen, X.; Shi, Y.; Gao, Y.; Lu, Z.; Shen, Q.; Li, L.; Pan, H. Accelerating Deep Neural Networks by Combining Block-Circulant Matrices and Low-Precision Weights. Electronics 2019, 8, 78.
Qin Z, Zhu D, Zhu X, Chen X, Shi Y, Gao Y, Lu Z, Shen Q, Li L, Pan H. Accelerating Deep Neural Networks by Combining Block-Circulant Matrices and Low-Precision Weights. Electronics. 2019; 8(1):78.
Qin, Zidi; Zhu, Di; Zhu, Xingwei; Chen, Xuan; Shi, Yinghuan; Gao, Yang; Lu, Zhonghai; Shen, Qinghong; Li, Li; Pan, Hongbing. 2019. "Accelerating Deep Neural Networks by Combining Block-Circulant Matrices and Low-Precision Weights." Electronics 8, no. 1: 78.
Show more citation formats
Show less citations formats
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.
[Return to top]
For more information on the journal statistics, click here
Multiple requests from the same IP address are counted as one view.