Conference In: orcid, cienciavitae

PositNN: Training Deep Neural Networks with Mixed Low-Precision Posit

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Gonçalo Raposo; Pedro Tomás; Nuno Roma — 2021 — IEEE

Key information

Authors:

Gonçalo Raposo (Gonçalo Eduardo Cascalho Raposo); Pedro Tomás (Pedro Filipe Zeferino Tomás); Nuno Roma (Nuno Filipe Valentim Roma)

Published in

June 2021

Abstract

Low-precision formats have proven to be an efficient way to reduce not only the memory footprint but also the hardware resources and power consumption of deep learning computations. Under this premise, the posit numerical format appears to be a highly viable substitute for the IEEE floating-point, but its application to neural networks training still requires further research. Some preliminary results have shown that 8-bit (and even smaller) posits may be used for inference and 16-bit for training, while maintaining the model accuracy. The presented research aims to evaluate the feasibility to train deep convolutional neural networks using posits. For such purpose, a software framework was developed to use simulated posits and quires in end-to-end training and inference. This implementation allows using any bit size, configuration, and even mixed precision, suitable for different precision requirements in various stages. The obtained results suggest that 8-bit posits can substitute 32-bit floats during training with no negative impact on the resulting loss and accuracy.

Publication details

Authors in the community:

Publication version

SMUR - Submitted manuscript under review

Publisher

IEEE

Link to the publisher's version

https://ieeexplore.ieee.org/document/9413919

Title of the publication container

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Location of the conference

Canada

Conference date start

06/06/2021

Conference date end

06/11/2021

ISBN

978-1-7281-7605-5

ISSN

1520-6149

DOI

10.1109/ICASSP39728.2021.9413919

Fields of Science and Technology (FOS)

electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering

Keywords

  • Posit numerical format
  • low-precision arithmetic
  • deep neural networks
  • training
  • inference

Publication language (ISO code)

eng - English

Rights type:

Open access