Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning

In this paper, we propose TECO, a multi-dimensional pruning framework to collaboratively prune the three dimensions (depth, width, and resolution) of convolutional neural networks (CNNs) for better execution efficiency on embedded hardware. In TECO, we first introduce a two-stage importance evaluati...

Full description

Saved in:
Bibliographic Details
Main Authors: Kong, Hao, Liu, Di, Luo, Xiangzhong, Huai, Shuo, Subramaniam, Ravi, Makaya, Christian, Lin, Qian, Liu, Weichen
Other Authors: School of Computer Science and Engineering
Format: Conference or Workshop Item
Language:English
Published: 2023
Subjects:
Online Access:https://hdl.handle.net/10356/167489
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-167489
record_format dspace
spelling sg-ntu-dr.10356-1674892023-12-15T02:06:54Z Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning Kong, Hao Liu, Di Luo, Xiangzhong Huai, Shuo Subramaniam, Ravi Makaya, Christian Lin, Qian Liu, Weichen School of Computer Science and Engineering 2023 60th ACM/IEEE Design Automation Conference (DAC) HP-NTU Digital Manufacturing Corporate Lab Engineering::Computer science and engineering Measurement Design Automation In this paper, we propose TECO, a multi-dimensional pruning framework to collaboratively prune the three dimensions (depth, width, and resolution) of convolutional neural networks (CNNs) for better execution efficiency on embedded hardware. In TECO, we first introduce a two-stage importance evaluation framework, which efficiently and comprehensively evaluates each pruning unit according to both the local importance inside each dimension and the global importance across different dimensions. Based on the evaluation framework, we present a heuristic pruning algorithm to progressively prune the three dimensions of CNNs towards the optimal trade-off between accuracy and efficiency. Experiments on multiple benchmarks validate the advantages of TECO over existing state-of-the-art (SOTA) approaches. The code and pre-trained models are available anonymously at https://github.com/ntuliuteam/Teco. Ministry of Education (MOE) Nanyang Technological University National Research Foundation (NRF) Submitted/Accepted version This study is partially supported under the RIE2020 Industry Alignment Fund – Industry Collaboration Projects (IAF-ICP) Funding Initiative, as well as cash and in-kind contribution from the industry partner, HP Inc., through the HP-NTU Digital Manufacturing Corporate Lab (I1801E0028). This work is also partially supported by the Ministry of Education, Singapore, under its Academic Research Fund Tier 2 (MOE2019-T2-1-071), and Nanyang Technological University, Singapore, under its NAP (M4082282). 2023-09-18T07:40:44Z 2023-09-18T07:40:44Z 2023 Conference Paper Kong, H., Liu, D., Luo, X., Huai, S., Subramaniam, R., Makaya, C., Lin, Q. & Liu, W. (2023). Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning. 2023 60th ACM/IEEE Design Automation Conference (DAC). https://dx.doi.org/10.1109/DAC56929.2023.10247965 979-8-3503-2348-1 https://hdl.handle.net/10356/167489 10.1109/DAC56929.2023.10247965 en I1801E0028 MOE2019-T2-1-071 NAP(M4082282) 10.21979/N9/BTNOJN © 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/DAC56929.2023.10247965. application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Engineering::Computer science and engineering
Measurement
Design Automation
spellingShingle Engineering::Computer science and engineering
Measurement
Design Automation
Kong, Hao
Liu, Di
Luo, Xiangzhong
Huai, Shuo
Subramaniam, Ravi
Makaya, Christian
Lin, Qian
Liu, Weichen
Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
description In this paper, we propose TECO, a multi-dimensional pruning framework to collaboratively prune the three dimensions (depth, width, and resolution) of convolutional neural networks (CNNs) for better execution efficiency on embedded hardware. In TECO, we first introduce a two-stage importance evaluation framework, which efficiently and comprehensively evaluates each pruning unit according to both the local importance inside each dimension and the global importance across different dimensions. Based on the evaluation framework, we present a heuristic pruning algorithm to progressively prune the three dimensions of CNNs towards the optimal trade-off between accuracy and efficiency. Experiments on multiple benchmarks validate the advantages of TECO over existing state-of-the-art (SOTA) approaches. The code and pre-trained models are available anonymously at https://github.com/ntuliuteam/Teco.
author2 School of Computer Science and Engineering
author_facet School of Computer Science and Engineering
Kong, Hao
Liu, Di
Luo, Xiangzhong
Huai, Shuo
Subramaniam, Ravi
Makaya, Christian
Lin, Qian
Liu, Weichen
format Conference or Workshop Item
author Kong, Hao
Liu, Di
Luo, Xiangzhong
Huai, Shuo
Subramaniam, Ravi
Makaya, Christian
Lin, Qian
Liu, Weichen
author_sort Kong, Hao
title Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
title_short Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
title_full Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
title_fullStr Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
title_full_unstemmed Towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
title_sort towards efficient convolutional neural network for embedded hardware via multi-dimensional pruning
publishDate 2023
url https://hdl.handle.net/10356/167489
_version_ 1787136482244296704