Enregistré dans:
Détails bibliographiques
Auteurs principaux: Jiang, Mengnan, Wang, Jingcun, Eldebiky, Amro, Yin, Xunzhao, Zhuo, Cheng, Lin, Ing-Chao, Zhang, Grace Li
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:https://arxiv.org/abs/2312.05875
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910335642894336
author Jiang, Mengnan
Wang, Jingcun
Eldebiky, Amro
Yin, Xunzhao
Zhuo, Cheng
Lin, Ing-Chao
Zhang, Grace Li
author_facet Jiang, Mengnan
Wang, Jingcun
Eldebiky, Amro
Yin, Xunzhao
Zhuo, Cheng
Lin, Ing-Chao
Zhang, Grace Li
contents Deep neural networks (DNNs) have demonstrated remarkable success in various fields. However, the large number of floating-point operations (FLOPs) in DNNs poses challenges for their deployment in resource-constrained applications, e.g., edge devices. To address the problem, pruning has been introduced to reduce the computational cost in executing DNNs. Previous pruning strategies are based on weight values, gradient values and activation outputs. Different from previous pruning solutions, in this paper, we propose a class-aware pruning technique to compress DNNs, which provides a novel perspective to reduce the computational cost of DNNs. In each iteration, the neural network training is modified to facilitate the class-aware pruning. Afterwards, the importance of filters with respect to the number of classes is evaluated. The filters that are only important for a few number of classes are removed. The neural network is then retrained to compensate for the incurred accuracy loss. The pruning iterations end until no filter can be removed anymore, indicating that the remaining filters are very important for many classes. This pruning technique outperforms previous pruning solutions in terms of accuracy, pruning ratio and the reduction of FLOPs. Experimental results confirm that this class-aware pruning technique can significantly reduce the number of weights and FLOPs, while maintaining a high inference accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2312_05875
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Class-Aware Pruning for Efficient Neural Networks
Jiang, Mengnan
Wang, Jingcun
Eldebiky, Amro
Yin, Xunzhao
Zhuo, Cheng
Lin, Ing-Chao
Zhang, Grace Li
Artificial Intelligence
Deep neural networks (DNNs) have demonstrated remarkable success in various fields. However, the large number of floating-point operations (FLOPs) in DNNs poses challenges for their deployment in resource-constrained applications, e.g., edge devices. To address the problem, pruning has been introduced to reduce the computational cost in executing DNNs. Previous pruning strategies are based on weight values, gradient values and activation outputs. Different from previous pruning solutions, in this paper, we propose a class-aware pruning technique to compress DNNs, which provides a novel perspective to reduce the computational cost of DNNs. In each iteration, the neural network training is modified to facilitate the class-aware pruning. Afterwards, the importance of filters with respect to the number of classes is evaluated. The filters that are only important for a few number of classes are removed. The neural network is then retrained to compensate for the incurred accuracy loss. The pruning iterations end until no filter can be removed anymore, indicating that the remaining filters are very important for many classes. This pruning technique outperforms previous pruning solutions in terms of accuracy, pruning ratio and the reduction of FLOPs. Experimental results confirm that this class-aware pruning technique can significantly reduce the number of weights and FLOPs, while maintaining a high inference accuracy.
title Class-Aware Pruning for Efficient Neural Networks
topic Artificial Intelligence
url https://arxiv.org/abs/2312.05875