Fast Exploration of Weight Sharing Opportunities for CNN Compression

TL;DR

This paper introduces an optimized exploration process to efficiently identify weight sharing opportunities in CNNs, significantly reducing design space exploration time without compromising compression quality.

Contribution

It presents a novel method to accelerate the exploration of weight sharing in CNN compression, addressing the scalability issue of existing DSE techniques.

Findings

Exploration time is drastically reduced.

Compression quality remains high.

Method applicable to various CNN architectures.

Abstract

The computational workload involved in Convolutional Neural Networks (CNNs) is typically out of reach for low-power embedded devices. There are a large number of approximation techniques to address this problem. These methods have hyper-parameters that need to be optimized for each CNNs using design space exploration (DSE). The goal of this work is to demonstrate that the DSE phase time can easily explode for state of the art CNN. We thus propose the use of an optimized exploration process to drastically reduce the exploration time without sacrificing the quality of the output.