A hardware-friendly pruning approach by exploiting local statistical pruning and fine grain pruning techniques

Chun Chi Chang, Chung Hsun Huang, Yuan Sun Chu

研究成果: Conference contribution

1 引文 斯高帕斯(Scopus)

摘要

Deep neural networks (DNN) have recently become a popular research topic and achieved well success in many signal processing tasks. However, when we deploy these neural networks on resource-limited hardware; e.g., edge devices, computation complexity and memory capability become great challenges. Many researches devote in compressing model parameters without significant accuracy loss, which is so-call pruning approach, so as to reduce the requirements of computation resource and memory space. However, pruned neural network (NN) model often shows heavy irregular sparsity between layers, convolution kernels, etc. Thus the utilizations of multiply-and-accumulate (MAC) array or processing element (PE) array could be low so that the inference time could not be reduced accordingly. In this paper, we propose a hardware-friendly pruning approach by exploiting local statistical pruning and fine-grain pruning techniques to possibly improve the utilizations of MAC array or PE array. Performance evaluations demonstrate that the performance of a NN-based super-resolution was kept good (i.e., > 37dB) with a high pruning ratio (i.e.,

原文English
主出版物標題2022 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2022
發行者Institute of Electrical and Electronics Engineers Inc.
ISBN(電子)9781665464345
DOIs
出版狀態Published - 2022
事件2022 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2022 - Yeosu, Korea, Republic of
持續時間: 2022 10月 262022 10月 28

出版系列

名字2022 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2022

Conference

Conference2022 IEEE International Conference on Consumer Electronics-Asia, ICCE-Asia 2022
國家/地區Korea, Republic of
城市Yeosu
期間22-10-2622-10-28

All Science Journal Classification (ASJC) codes

  • 電腦科學應用
  • 電氣與電子工程
  • 媒體技術
  • 儀器

引用此