Fast Algorithms for Convolutional Neural Networks

Deep convolutional neural networks take GPU-days of computation to train on large data sets. Pedestrian detection for self driving cars requires very low latency. Image recognition for mobile phones is constrained by limited processing resources. The success of convolutional neural networks in these...

Full description

Saved in:
Bibliographic Details
Published in2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) pp. 4013 - 4021
Main Authors Lavin, Andrew, Gray, Scott
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2016
Subjects
Complexity theory
Convolution
Image recognition
Laplace equations
Neural networks
Transforms
Two dimensional displays
Online AccessGet full text
ISSN1063-6919
DOI10.1109/CVPR.2016.435

Cover

More Information
Summary:Deep convolutional neural networks take GPU-days of computation to train on large data sets. Pedestrian detection for self driving cars requires very low latency. Image recognition for mobile phones is constrained by limited processing resources. The success of convolutional neural networks in these situations is limited by how fast we can compute them. Conventional FFT based convolution is fast for large filters, but state of the art convolutional neural networks use small, 3 3 filters. We introduce a new class of fast algorithms for convolutional neural networks using Winograd's minimal filtering algorithms. The algorithms compute minimal complexity convolution over small tiles, which makes them fast with small filters and small batch sizes. We benchmark a GPU implementation of our algorithm with the VGG network and show state of the art throughput at batch sizes from 1 to 64.
ISSN:1063-6919
DOI:10.1109/CVPR.2016.435