Convolutional neural networks compression with low rank and sparse tensor decompositions

TL;DR

This paper introduces a neural network compression technique using tensor decompositions that combine low-rank and sparse components, significantly reducing model size and increasing speed for convolutional layers.

Contribution

It proposes a novel tensor decomposition method for CNN compression that leverages low-rank and sparse approximations to eliminate different redundancies, improving efficiency.

Findings

Achieved up to 3.5x CPU speedup for convolutional layers.

Reduced layer size by up to 11x in ResNet50.

Demonstrated effectiveness on image classification tasks.

Abstract

Convolutional neural networks show outstanding results in a variety of computer vision tasks. However, a neural network architecture design usually faces a trade-off between model performance and computational/memory complexity. For some real-world applications, it is crucial to develop models, which can be fast and light enough to run on edge systems and mobile devices. However, many modern architectures that demonstrate good performance don't satisfy inference time and storage limitation requirements. Thus, arises a problem of neural network compression to obtain a smaller and faster model, which is on par with the initial one. In this work, we consider a neural network compression method based on tensor decompositions. Namely, we propose to approximate the convolutional layer weight with a tensor, which can be represented as a sum of low-rank and sparse components. The motivation for such approximation is based on the assumption that low-rank and sparse terms allow eliminating two different types of redundancy and thus yield a better compression rate. An efficient CPU implementation for the proposed method has been developed. Our algorithm has demonstrated up to 3.5x CPU layer speedup and 11x layer size reduction when compressing Resnet50 architecture for the image classification task.