Multi-precision convolutional neural networks on heterogeneous hardware

Sam Amiri, Mohammad Hosseinabady, Simon McIntosh-Smith, Jose Nunez-Yanez

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

13 Citations (Scopus)
488 Downloads (Pure)


Fully binarised convolutional neural networks (CNNs) deliver very high inference performance using singlebit weights and activations, together with XNOR type operators for the kernel convolutions. Current research shows that full binarisation results in a degradation of accuracy and different
approaches to tackle this issue are being investigated such as using more complex models as accuracy reduces. This paper proposes an alternative based on a multi-precision CNN framework
that combines a binarised and a floating point CNN in a pipeline configuration deployed on heterogeneous hardware. The binarised CNN is mapped onto an FPGA device and used to perform inference over the whole input set while the floating point network is mapped onto a CPU device and performs reinference only when the classification confidence level is low. A light-weight confidence mechanism enables a flexible trade-off between accuracy and throughput. To demonstrate the concept, we choose a Zynq 7020 device as the hardware target and show
that the multi-precision network is able to increase the BNN accuracy from 78.5% to 82.5% and the CPU inference speed from 29.68 to 90.82 images/sec.
Original languageEnglish
Title of host publicationDesign, Automation & Test in Europe Conference & Exhibition (DATE), 2018
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Number of pages6
ISBN (Electronic)9783981926309, 9783981926316
Publication statusE-pub ahead of print - 23 Apr 2018
EventDATE 2018 - Dresden, Germany
Duration: 19 Mar 201823 Mar 2018

Publication series

ISSN (Print)1558-1101


ConferenceDATE 2018


Dive into the research topics of 'Multi-precision convolutional neural networks on heterogeneous hardware'. Together they form a unique fingerprint.

Cite this