Soft Masking for Cost-Constrained Channel Pruning

TL;DR

This paper introduces SMCP, a novel soft masking approach for channel pruning in CNNs that allows pruned channels to recover, improving accuracy under cost constraints on ImageNet and PASCAL VOC.

Contribution

We propose a soft mask re-parameterization method for channel pruning that enables channels to adaptively return, addressing accuracy loss in traditional pruning methods.

Findings

Outperforms prior channel pruning methods on ImageNet.

Achieves better accuracy with fewer channels.

Effectively balances model size and performance.

Abstract

Structured channel pruning has been shown to significantly accelerate inference time for convolution neural networks (CNNs) on modern hardware, with a relatively minor loss of network accuracy. Recent works permanently zero these channels during training, which we observe to significantly hamper final accuracy, particularly as the fraction of the network being pruned increases. We propose Soft Masking for cost-constrained Channel Pruning (SMCP) to allow pruned channels to adaptively return to the network while simultaneously pruning towards a target cost constraint. By adding a soft mask re-parameterization of the weights and channel pruning from the perspective of removing input channels, we allow gradient updates to previously pruned channels and the opportunity for the channels to later return to the network. We then formulate input channel pruning as a global resource allocation problem. Our method outperforms prior works on both the ImageNet classification and PASCAL VOC detection datasets.