COP: customized correlation-based Filter level pruning method for deep CNN compression

doi:10.1016/j.neucom.2021.08.098

COP: customized correlation-based Filter level pruning method for deep CNN compression

2021-11-01

14

PRE

AI

Xiaofei He (Xiaofei He)

DOI：10.1016/j.neucom.2021.08.098

原文链接

原文求助

摘要

En 中文

As deep CNNs get larger, it becomes more challenging to deploy them on resource-restricted mobile devices. Filter-level pruning is one of the most popular methods to compress deep models for mobile deployment. It prunes unimportant filters in the pre-trained CNN to reduce its storage and computational cost, yielding a smaller and more efficient model. Though having made some progress, most filter-level pruning methods still suffer from at least one of the following two problems: 1) High redundancy: some methods pick out unimportant filters without considering their correlations, and thus many highly correlated filters are not pruned, yielding a model still with high redundancy. 2) Sub-optimal: all existing pruning methods achieve sub-optimal pruning plan because they neglect that a high parameters reduc-tion ratio (PRR) does not always mean a high FLOPs reduction ratio (FRR). In this paper, we propose our customized correlation-based pruning (COP) to solve these problems. In particular, we observe redundant filters through their correlations. Moreover, to achieve the optimal pruning plan, PRR and FRR are consid-ered when evaluating filters' importance. Besides, we also propose a new pruning pipeline, which improves the accuracy of the pruned model. Extensive experiments show that our proposed method has outperformed the state-of-the-art on several popular architectures and datasets. (c) 2021 Elsevier B.V. All rights reserved.

Keyword:

Model compression

Model acceleration

Convolutional neural network

Neural network pruning

COP: customized correlation-based Filter level pruning method for deep CNN compression

摘要

期刊

机构

引用论文