Quantization of deep neural networks for accumulator-constrained processors

Research output: Contribution to journalConference articleAcademicpeer-review

Abstract

We introduce an Artificial Neural Network (ANN) quantization methodology for platforms without wide accumulation registers. This enables fixed-point model deployment on embedded compute platforms that are not specifically designed for large kernel computations (i.e. accumulator-constrained processors). We formulate the quantization problem as a function of accumulator size, and aim to maximize the model accuracy by maximizing bit width of input data and weights. To reduce the number of configurations to consider, only solutions that fully utilize the available accumulator bits are being tested. We demonstrate that 16-bit accumulators are able to obtain a classification accuracy within 1% of the floating-point baselines on the CIFAR-10 and ILSVRC2012 image classification benchmarks. Additionally, a near-optimal 2 ×  speedup is obtained on an ARM processor, by exploiting 16-bit accumulators for image classification on the All-CNN-C and AlexNet networks.
LanguageEnglish
JournalMicroprocessors and Microsystems
DOIs
StateE-pub ahead of print - 31 Aug 2019

Fingerprint

Image classification
ARM processors
Neural networks
Deep neural networks

Cite this

@article{de8a7c70cba94bde89e92c0fb10f1bc6,
title = "Quantization of deep neural networks for accumulator-constrained processors",
abstract = "We introduce an Artificial Neural Network (ANN) quantization methodology for platforms without wide accumulation registers. This enables fixed-point model deployment on embedded compute platforms that are not specifically designed for large kernel computations (i.e. accumulator-constrained processors). We formulate the quantization problem as a function of accumulator size, and aim to maximize the model accuracy by maximizing bit width of input data and weights. To reduce the number of configurations to consider, only solutions that fully utilize the available accumulator bits are being tested. We demonstrate that 16-bit accumulators are able to obtain a classification accuracy within 1{\%} of the floating-point baselines on the CIFAR-10 and ILSVRC2012 image classification benchmarks. Additionally, a near-optimal 2 ×  speedup is obtained on an ARM processor, by exploiting 16-bit accumulators for image classification on the All-CNN-C and AlexNet networks.",
author = "{de Bruin}, Barry and Zoran Zivkovic and Henk Corporaal",
year = "2019",
month = "8",
day = "31",
doi = "10.1016/j.micpro.2019.102872",
language = "English",
journal = "Microprocessors and Microsystems",
issn = "0141-9331",
publisher = "Elsevier",

}

Quantization of deep neural networks for accumulator-constrained processors. / de Bruin, Barry; Zivkovic, Zoran; Corporaal, Henk.

In: Microprocessors and Microsystems, 31.08.2019.

Research output: Contribution to journalConference articleAcademicpeer-review

TY - JOUR

T1 - Quantization of deep neural networks for accumulator-constrained processors

AU - de Bruin,Barry

AU - Zivkovic,Zoran

AU - Corporaal,Henk

PY - 2019/8/31

Y1 - 2019/8/31

N2 - We introduce an Artificial Neural Network (ANN) quantization methodology for platforms without wide accumulation registers. This enables fixed-point model deployment on embedded compute platforms that are not specifically designed for large kernel computations (i.e. accumulator-constrained processors). We formulate the quantization problem as a function of accumulator size, and aim to maximize the model accuracy by maximizing bit width of input data and weights. To reduce the number of configurations to consider, only solutions that fully utilize the available accumulator bits are being tested. We demonstrate that 16-bit accumulators are able to obtain a classification accuracy within 1% of the floating-point baselines on the CIFAR-10 and ILSVRC2012 image classification benchmarks. Additionally, a near-optimal 2 ×  speedup is obtained on an ARM processor, by exploiting 16-bit accumulators for image classification on the All-CNN-C and AlexNet networks.

AB - We introduce an Artificial Neural Network (ANN) quantization methodology for platforms without wide accumulation registers. This enables fixed-point model deployment on embedded compute platforms that are not specifically designed for large kernel computations (i.e. accumulator-constrained processors). We formulate the quantization problem as a function of accumulator size, and aim to maximize the model accuracy by maximizing bit width of input data and weights. To reduce the number of configurations to consider, only solutions that fully utilize the available accumulator bits are being tested. We demonstrate that 16-bit accumulators are able to obtain a classification accuracy within 1% of the floating-point baselines on the CIFAR-10 and ILSVRC2012 image classification benchmarks. Additionally, a near-optimal 2 ×  speedup is obtained on an ARM processor, by exploiting 16-bit accumulators for image classification on the All-CNN-C and AlexNet networks.

U2 - 10.1016/j.micpro.2019.102872

DO - 10.1016/j.micpro.2019.102872

M3 - Conference article

JO - Microprocessors and Microsystems

T2 - Microprocessors and Microsystems

JF - Microprocessors and Microsystems

SN - 0141-9331

ER -