An optimization method for pruning rates of each layer in CNN based on the GA-SMSM
Parameter pruning is one of the primary methods for compressing CNN models, aiming to reduce redundant parameters, the complexity of time and space, and the calculation resources of the network, all while ensuring minimal loss in the network’s performance. Currently, most existing parameter pruning...
Gespeichert in:
| Veröffentlicht in: | Memetic computing 2024-03, Vol.16 (1), p.45-54 |
|---|---|
| Hauptverfasser: | , , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 54 |
|---|---|
| container_issue | 1 |
| container_start_page | 45 |
| container_title | Memetic computing |
| container_volume | 16 |
| creator | Dong, Xiaoyu Yan, Pinshuai Wang, Mengfei Li, Binqi Song, Yuantao |
| description | Parameter pruning is one of the primary methods for compressing CNN models, aiming to reduce redundant parameters, the complexity of time and space, and the calculation resources of the network, all while ensuring minimal loss in the network’s performance. Currently, most existing parameter pruning methods adopt equal pruning rates across all layers. Different from previous methods, this paper focuses on the optimal combination of each layer’s pruning rates within a given pruning rate of the whole model. Genetic algorithm is used to determine the pruning rate for each layer. It’s worth noting that while the pruning rate for individual layers may vary, the average pruning rate across all layers does not exceed the given pruning rate. Experimental validation is conducted on CIFAR10 and ImageNet ILSVRC2012 datasets using VGGNet and ResNet architectures. The results show that the accuracy loss and the FLOPs of the pruned model using our method are superior to those pruned using previous methods. |
| doi_str_mv | 10.1007/s12293-023-00402-2 |
| format | Article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_2956785825</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2956785825</sourcerecordid><originalsourceid>FETCH-LOGICAL-c270t-159fc2085822efb698ac95be7a732c6643e8c54c8ff96ac7c22c075dfc8eef3</originalsourceid><addsrcrecordid>eNp9kMFKAzEQhoMoWGpfwFPA82p2drNJjqVoFdoK1ntI00m7pd2syfZQn97UFb05ECaH__sHPkJuc3afMyYeYg6gioxBeqxkkMEFGeSy4pkCBZe_f1lek1GMO5amACHLfEDexg31bVcf6k_T1b6hB-y2fk2dD7QNx6ZuNjSYDiP1jqKxW7o3Jwy0buhksaArE3FNE9ZtkU7H2XK-nN-QK2f2EUc_e0iWT4_vk-ds9jp9mYxnmQXBuiznyllgkksAdKtKSWMVX6EwogBbVWWB0vLSSudUZaywAJYJvnZWIrpiSO761jb4jyPGTu_8MTTpoAbFK3Hu5SkFfcoGH2NAp9tQH0w46Zzpszzdy9NJnv6WpyFBRQ_FFG42GP6q_6G-AJ99cGE</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2956785825</pqid></control><display><type>article</type><title>An optimization method for pruning rates of each layer in CNN based on the GA-SMSM</title><source>SpringerLink Journals</source><creator>Dong, Xiaoyu ; Yan, Pinshuai ; Wang, Mengfei ; Li, Binqi ; Song, Yuantao</creator><creatorcontrib>Dong, Xiaoyu ; Yan, Pinshuai ; Wang, Mengfei ; Li, Binqi ; Song, Yuantao</creatorcontrib><description>Parameter pruning is one of the primary methods for compressing CNN models, aiming to reduce redundant parameters, the complexity of time and space, and the calculation resources of the network, all while ensuring minimal loss in the network’s performance. Currently, most existing parameter pruning methods adopt equal pruning rates across all layers. Different from previous methods, this paper focuses on the optimal combination of each layer’s pruning rates within a given pruning rate of the whole model. Genetic algorithm is used to determine the pruning rate for each layer. It’s worth noting that while the pruning rate for individual layers may vary, the average pruning rate across all layers does not exceed the given pruning rate. Experimental validation is conducted on CIFAR10 and ImageNet ILSVRC2012 datasets using VGGNet and ResNet architectures. The results show that the accuracy loss and the FLOPs of the pruned model using our method are superior to those pruned using previous methods.</description><identifier>ISSN: 1865-9284</identifier><identifier>EISSN: 1865-9292</identifier><identifier>DOI: 10.1007/s12293-023-00402-2</identifier><language>eng</language><publisher>Berlin/Heidelberg: Springer Berlin Heidelberg</publisher><subject>Applications of Mathematics ; Artificial Intelligence ; Bioinformatics ; Complex Systems ; Control ; Engineering ; Genetic algorithms ; Mathematical and Computational Engineering ; Mathematical models ; Mechatronics ; Parameters ; Regular Research paper ; Robotics</subject><ispartof>Memetic computing, 2024-03, Vol.16 (1), p.45-54</ispartof><rights>The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c270t-159fc2085822efb698ac95be7a732c6643e8c54c8ff96ac7c22c075dfc8eef3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/s12293-023-00402-2$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/s12293-023-00402-2$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>314,776,780,27903,27904,41467,42536,51297</link.rule.ids></links><search><creatorcontrib>Dong, Xiaoyu</creatorcontrib><creatorcontrib>Yan, Pinshuai</creatorcontrib><creatorcontrib>Wang, Mengfei</creatorcontrib><creatorcontrib>Li, Binqi</creatorcontrib><creatorcontrib>Song, Yuantao</creatorcontrib><title>An optimization method for pruning rates of each layer in CNN based on the GA-SMSM</title><title>Memetic computing</title><addtitle>Memetic Comp</addtitle><description>Parameter pruning is one of the primary methods for compressing CNN models, aiming to reduce redundant parameters, the complexity of time and space, and the calculation resources of the network, all while ensuring minimal loss in the network’s performance. Currently, most existing parameter pruning methods adopt equal pruning rates across all layers. Different from previous methods, this paper focuses on the optimal combination of each layer’s pruning rates within a given pruning rate of the whole model. Genetic algorithm is used to determine the pruning rate for each layer. It’s worth noting that while the pruning rate for individual layers may vary, the average pruning rate across all layers does not exceed the given pruning rate. Experimental validation is conducted on CIFAR10 and ImageNet ILSVRC2012 datasets using VGGNet and ResNet architectures. The results show that the accuracy loss and the FLOPs of the pruned model using our method are superior to those pruned using previous methods.</description><subject>Applications of Mathematics</subject><subject>Artificial Intelligence</subject><subject>Bioinformatics</subject><subject>Complex Systems</subject><subject>Control</subject><subject>Engineering</subject><subject>Genetic algorithms</subject><subject>Mathematical and Computational Engineering</subject><subject>Mathematical models</subject><subject>Mechatronics</subject><subject>Parameters</subject><subject>Regular Research paper</subject><subject>Robotics</subject><issn>1865-9284</issn><issn>1865-9292</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNp9kMFKAzEQhoMoWGpfwFPA82p2drNJjqVoFdoK1ntI00m7pd2syfZQn97UFb05ECaH__sHPkJuc3afMyYeYg6gioxBeqxkkMEFGeSy4pkCBZe_f1lek1GMO5amACHLfEDexg31bVcf6k_T1b6hB-y2fk2dD7QNx6ZuNjSYDiP1jqKxW7o3Jwy0buhksaArE3FNE9ZtkU7H2XK-nN-QK2f2EUc_e0iWT4_vk-ds9jp9mYxnmQXBuiznyllgkksAdKtKSWMVX6EwogBbVWWB0vLSSudUZaywAJYJvnZWIrpiSO761jb4jyPGTu_8MTTpoAbFK3Hu5SkFfcoGH2NAp9tQH0w46Zzpszzdy9NJnv6WpyFBRQ_FFG42GP6q_6G-AJ99cGE</recordid><startdate>20240301</startdate><enddate>20240301</enddate><creator>Dong, Xiaoyu</creator><creator>Yan, Pinshuai</creator><creator>Wang, Mengfei</creator><creator>Li, Binqi</creator><creator>Song, Yuantao</creator><general>Springer Berlin Heidelberg</general><general>Springer Nature B.V</general><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>20240301</creationdate><title>An optimization method for pruning rates of each layer in CNN based on the GA-SMSM</title><author>Dong, Xiaoyu ; Yan, Pinshuai ; Wang, Mengfei ; Li, Binqi ; Song, Yuantao</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c270t-159fc2085822efb698ac95be7a732c6643e8c54c8ff96ac7c22c075dfc8eef3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Applications of Mathematics</topic><topic>Artificial Intelligence</topic><topic>Bioinformatics</topic><topic>Complex Systems</topic><topic>Control</topic><topic>Engineering</topic><topic>Genetic algorithms</topic><topic>Mathematical and Computational Engineering</topic><topic>Mathematical models</topic><topic>Mechatronics</topic><topic>Parameters</topic><topic>Regular Research paper</topic><topic>Robotics</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Dong, Xiaoyu</creatorcontrib><creatorcontrib>Yan, Pinshuai</creatorcontrib><creatorcontrib>Wang, Mengfei</creatorcontrib><creatorcontrib>Li, Binqi</creatorcontrib><creatorcontrib>Song, Yuantao</creatorcontrib><collection>CrossRef</collection><jtitle>Memetic computing</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Dong, Xiaoyu</au><au>Yan, Pinshuai</au><au>Wang, Mengfei</au><au>Li, Binqi</au><au>Song, Yuantao</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>An optimization method for pruning rates of each layer in CNN based on the GA-SMSM</atitle><jtitle>Memetic computing</jtitle><stitle>Memetic Comp</stitle><date>2024-03-01</date><risdate>2024</risdate><volume>16</volume><issue>1</issue><spage>45</spage><epage>54</epage><pages>45-54</pages><issn>1865-9284</issn><eissn>1865-9292</eissn><abstract>Parameter pruning is one of the primary methods for compressing CNN models, aiming to reduce redundant parameters, the complexity of time and space, and the calculation resources of the network, all while ensuring minimal loss in the network’s performance. Currently, most existing parameter pruning methods adopt equal pruning rates across all layers. Different from previous methods, this paper focuses on the optimal combination of each layer’s pruning rates within a given pruning rate of the whole model. Genetic algorithm is used to determine the pruning rate for each layer. It’s worth noting that while the pruning rate for individual layers may vary, the average pruning rate across all layers does not exceed the given pruning rate. Experimental validation is conducted on CIFAR10 and ImageNet ILSVRC2012 datasets using VGGNet and ResNet architectures. The results show that the accuracy loss and the FLOPs of the pruned model using our method are superior to those pruned using previous methods.</abstract><cop>Berlin/Heidelberg</cop><pub>Springer Berlin Heidelberg</pub><doi>10.1007/s12293-023-00402-2</doi><tpages>10</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1865-9284 |
| ispartof | Memetic computing, 2024-03, Vol.16 (1), p.45-54 |
| issn | 1865-9284 1865-9292 |
| language | eng |
| recordid | cdi_proquest_journals_2956785825 |
| source | SpringerLink Journals |
| subjects | Applications of Mathematics Artificial Intelligence Bioinformatics Complex Systems Control Engineering Genetic algorithms Mathematical and Computational Engineering Mathematical models Mechatronics Parameters Regular Research paper Robotics |
| title | An optimization method for pruning rates of each layer in CNN based on the GA-SMSM |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-25T11%3A10%3A19IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=An%20optimization%20method%20for%20pruning%20rates%20of%20each%20layer%20in%20CNN%20based%20on%20the%20GA-SMSM&rft.jtitle=Memetic%20computing&rft.au=Dong,%20Xiaoyu&rft.date=2024-03-01&rft.volume=16&rft.issue=1&rft.spage=45&rft.epage=54&rft.pages=45-54&rft.issn=1865-9284&rft.eissn=1865-9292&rft_id=info:doi/10.1007/s12293-023-00402-2&rft_dat=%3Cproquest_cross%3E2956785825%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2956785825&rft_id=info:pmid/&rfr_iscdi=true |