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

Authors: de Bruin, Barry and Zivkovic, Zoran and Corporaal, Henk

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 2x speedup is obtained on an ARM processor, by exploiting 16 bit accumulators for image classification on the All-CNN-C and AlexNet networks.

[BibTeX] [ DOI]
@article{Bruin2020,
author = {de Bruin, Barry and Zivkovic, Zoran and Corporaal, Henk},
doi = {10.1016/j.micpro.2019.102872},
issn = {01419331},
journal = {Microprocessors and Microsystems},
keywords = {Convolutional neural networks,Efficient inference,Fixed-point,Narrow accumulators,Quantization},
pages = {102872},
publisher = {Elsevier B.V.},
title = {{Quantization of deep neural networks for accumulator-constrained processors}},
volume = {72},
url = {https://arxiv.org/pdf/2004.11783.pdf},
year = {2020}
}