Structured Probabilistic Pruning for Convolutional Neural Network Acceleration

TL;DR

This paper introduces Structured Probabilistic Pruning, a novel method for CNN acceleration that probabilistically prunes weights, achieving significant speedups with minimal accuracy loss and applicability to various network architectures.

Contribution

The paper presents a probabilistic pruning approach that improves CNN acceleration by dynamically adjusting pruning probabilities, outperforming deterministic methods.

Findings

Achieves 4x speedup with only 0.3% top-5 accuracy loss on AlexNet.

Attains 0.8% accuracy loss with 4x speedup on VGG-16.

Successfully accelerates ResNet-50 with 2x speedup and minimal accuracy loss.

Abstract

In this paper, we propose a novel progressive parameter pruning method for Convolutional Neural Network acceleration, named Structured Probabilistic Pruning (SPP), which effectively prunes weights of convolutional layers in a probabilistic manner. Unlike existing deterministic pruning approaches, where unimportant weights are permanently eliminated, SPP introduces a pruning probability for each weight, and pruning is guided by sampling from the pruning probabilities. A mechanism is designed to increase and decrease pruning probabilities based on importance criteria in the training process. Experiments show that, with 4x speedup, SPP can accelerate AlexNet with only 0.3% loss of top-5 accuracy and VGG-16 with 0.8% loss of top-5 accuracy in ImageNet classification. Moreover, SPP can be directly applied to accelerate multi-branch CNN networks, such as ResNet, without specific adaptations. Our 2x speedup ResNet-50 only suffers 0.8% loss of top-5 accuracy on ImageNet. We further show the effectiveness of SPP on transfer learning tasks.