Speeding up Convolutional Neural Networks with Low Rank Expansions

TL;DR

This paper introduces two simple, architecture-agnostic methods for accelerating convolutional neural networks by exploiting filter redundancy, achieving significant speedups with minimal accuracy loss.

Contribution

It presents low rank expansion schemes for convolutional layers that can be easily integrated into existing frameworks to improve speed without sacrificing accuracy.

Findings

Achieved up to 4.5x speedup with less than 1% accuracy drop.

Demonstrated effectiveness on scene text recognition network.

Maintained state-of-the-art performance on benchmarks.

Abstract

The focus of this paper is speeding up the evaluation of convolutional neural networks. While delivering impressive results across a range of computer vision and machine learning tasks, these networks are computationally demanding, limiting their deployability. Convolutional layers generally consume the bulk of the processing time, and so in this work we present two simple schemes for drastically speeding up these layers. This is achieved by exploiting cross-channel or filter redundancy to construct a low rank basis of filters that are rank-1 in the spatial domain. Our methods are architecture agnostic, and can be easily applied to existing CPU and GPU convolutional frameworks for tuneable speedup performance. We demonstrate this with a real world network designed for scene text character recognition, showing a possible 2.5x speedup with no loss in accuracy, and 4.5x speedup with less than 1% drop in accuracy, still achieving state-of-the-art on standard benchmarks.