Article: Learning the number of filters in convolutional neural networks Journal: International Journal of Bio-Inspired Computation (IJBIC) 2021 Vol.17 No.2 pp.75 - 84 Abstract: Convolutional networks bring the performance of many computer vision tasks to unprecedented heights, but at the cost of enormous computation load. To reduce this cost, many model compression tasks have been proposed by eliminating insignificant model structures. For example, convolution filters with small absolute weights are pruned and then fine-tuned to restore reasonable accuracy. However, most of these works rely on pre-trained models without specific analysis of the changes in filters during the training process, resulting in sizable model retraining costs. Different from previous works, we interpret the change of filter behaviour during training from the associated angle, and propose a novel filter pruning method utilising the change rule, which can remove filters with similar functions later in training. According to this strategy, not only can we achieve model compression without fine-tuning, but we can also find a novel perspective to interpret the changing behaviour of the filter during training. Moreover, our approach has been proved to be effective for many advanced CNN architectures. Inderscience Publishers - linking academia, business and industry through research

Title: Learning the number of filters in convolutional neural networks

Authors: Jue Li; Feng Cao; Honghong Cheng; Yuhua Qian

Addresses: Institute of Big Data Science and Industry, School of Computer and Information Technology, Shanxi University, Taiyuan, China ' School of Computer and Information Technology, Shanxi University, Taiyuan, China ' Institute of Big Data Science and Industry, Shanxi University, Taiyuan, China ' Institute of Big Data Science and Industry, School of Computer and Information Technology, Shanxi University, Taiyuan, China

Abstract: Convolutional networks bring the performance of many computer vision tasks to unprecedented heights, but at the cost of enormous computation load. To reduce this cost, many model compression tasks have been proposed by eliminating insignificant model structures. For example, convolution filters with small absolute weights are pruned and then fine-tuned to restore reasonable accuracy. However, most of these works rely on pre-trained models without specific analysis of the changes in filters during the training process, resulting in sizable model retraining costs. Different from previous works, we interpret the change of filter behaviour during training from the associated angle, and propose a novel filter pruning method utilising the change rule, which can remove filters with similar functions later in training. According to this strategy, not only can we achieve model compression without fine-tuning, but we can also find a novel perspective to interpret the changing behaviour of the filter during training. Moreover, our approach has been proved to be effective for many advanced CNN architectures.

Keywords: model compress; filter pruning; filter correlation; filter behaviour interpretable.

DOI: 10.1504/IJBIC.2021.114101

International Journal of Bio-Inspired Computation, 2021 Vol.17 No.2, pp.75 - 84

Received: 22 Jul 2020
Accepted: 25 Sep 2020
Published online: 08 Apr 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Learning the number of filters in convolutional neural networks

Keep up-to-date