Dynamic Network Surgery for Efficient DNNs

TL;DR

This paper introduces dynamic network surgery, a novel method for compressing deep neural networks by on-the-fly pruning and splicing, significantly reducing model size without accuracy loss.

Contribution

It proposes a new network compression technique that integrates connection pruning and splicing for continual network maintenance, outperforming previous pruning methods.

Findings

Compresses LeNet-5 by 108x without accuracy loss

Reduces AlexNet size by 17.7x without accuracy loss

Outperforms recent pruning methods significantly

Abstract

Deep learning has become a ubiquitous technology to improve machine intelligence. However, most of the existing deep models are structurally very complex, making them difficult to be deployed on the mobile platforms with limited computational power. In this paper, we propose a novel network compression method called dynamic network surgery, which can remarkably reduce the network complexity by making on-the-fly connection pruning. Unlike the previous methods which accomplish this task in a greedy way, we properly incorporate connection splicing into the whole process to avoid incorrect pruning and make it as a continual network maintenance. The effectiveness of our method is proved with experiments. Without any accuracy loss, our method can efficiently compress the number of parameters in LeNet-5 and AlexNet by a factor of $\bm{108}\times$ and $\bm{17.7}\times$ respectively, proving that it outperforms the recent pruning method by considerable margins. Code and some models are available at https://github.com/yiwenguo/Dynamic-Network-Surgery.