{"id":"https://openalex.org/W4404035086","doi":"https://doi.org/10.1109/tcsi.2024.3485192","title":"Trio-ViT: Post-Training Quantization and Acceleration for Softmax-Free Efficient Vision Transformer","display_name":"Trio-ViT: Post-Training Quantization and Acceleration for Softmax-Free Efficient Vision Transformer","publication_year":2024,"publication_date":"2024-11-04","ids":{"openalex":"https://openalex.org/W4404035086","doi":"https://doi.org/10.1109/tcsi.2024.3485192"},"language":"en","primary_location":{"id":"doi:10.1109/tcsi.2024.3485192","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsi.2024.3485192","pdf_url":null,"source":{"id":"https://openalex.org/S116977442","display_name":"IEEE Transactions on Circuits and Systems I Regular Papers","issn_l":"1549-8328","issn":["1549-8328","1558-0806"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Circuits and Systems I: Regular Papers","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Huihong Shi","orcid":"https://orcid.org/0000-0002-7845-0154"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huihong Shi","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0002-7845-0154","affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090170161","display_name":"Haikuo Shao","orcid":"https://orcid.org/0009-0008-6965-3436"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haikuo Shao","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0009-0008-6965-3436","affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101895317","display_name":"Wendong Mao","orcid":"https://orcid.org/0000-0002-1061-7165"},"institutions":[{"id":"https://openalex.org/I157773358","display_name":"Sun Yat-sen University","ror":"https://ror.org/0064kty71","country_code":"CN","type":"education","lineage":["https://openalex.org/I157773358"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wendong Mao","raw_affiliation_strings":["School of Integrated Circuits, Sun Yat-sen University, Shenzhen, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Integrated Circuits, Sun Yat-sen University, Shenzhen, China","institution_ids":["https://openalex.org/I157773358"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100696999","display_name":"Zhongfeng Wang","orcid":"https://orcid.org/0000-0002-7227-4786"},"institutions":[{"id":"https://openalex.org/I881766915","display_name":"Nanjing University","ror":"https://ror.org/01rxvg760","country_code":"CN","type":"education","lineage":["https://openalex.org/I881766915"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhongfeng Wang","raw_affiliation_strings":["School of Electronic Science and Engineering, Nanjing University, Nanjing, China"],"raw_orcid":"https://orcid.org/0000-0002-7227-4786","affiliations":[{"raw_affiliation_string":"School of Electronic Science and Engineering, Nanjing University, Nanjing, China","institution_ids":["https://openalex.org/I881766915"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":{"value":2495,"currency":"USD","value_usd":2495},"apc_paid":null,"fwci":1.286,"has_fulltext":false,"cited_by_count":11,"citation_normalized_percentile":{"value":0.78928261,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":98,"max":99},"biblio":{"volume":"72","issue":"3","first_page":"1296","last_page":"1307"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11992","display_name":"CCD and CMOS Imaging Sensors","score":0.9980999827384949,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11992","display_name":"CCD and CMOS Imaging Sensors","score":0.9980999827384949,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12389","display_name":"Infrared Target Detection Methodologies","score":0.9818000197410583,"subfield":{"id":"https://openalex.org/subfields/2202","display_name":"Aerospace Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T13114","display_name":"Image Processing Techniques and Applications","score":0.9732000231742859,"subfield":{"id":"https://openalex.org/subfields/2214","display_name":"Media Technology"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/softmax-function","display_name":"Softmax function","score":0.7351433634757996},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.616065263748169},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5420028567314148},{"id":"https://openalex.org/keywords/quantization","display_name":"Quantization (signal processing)","score":0.5030285716056824},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4194572865962982},{"id":"https://openalex.org/keywords/electrical-engineering","display_name":"Electrical engineering","score":0.3559657335281372},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.3514842987060547},{"id":"https://openalex.org/keywords/computer-vision","display_name":"Computer vision","score":0.3174351453781128},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.2959122061729431},{"id":"https://openalex.org/keywords/voltage","display_name":"Voltage","score":0.15459471940994263},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.11746394634246826}],"concepts":[{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.7351433634757996},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.616065263748169},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5420028567314148},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.5030285716056824},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4194572865962982},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.3559657335281372},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3514842987060547},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3174351453781128},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.2959122061729431},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.15459471940994263},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.11746394634246826}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tcsi.2024.3485192","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tcsi.2024.3485192","pdf_url":null,"source":{"id":"https://openalex.org/S116977442","display_name":"IEEE Transactions on Circuits and Systems I Regular Papers","issn_l":"1549-8328","issn":["1549-8328","1558-0806"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Circuits and Systems I: Regular Papers","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","display_name":"Affordable and clean energy","score":0.8100000023841858}],"awards":[{"id":"https://openalex.org/G6422954184","display_name":null,"funder_award_id":"2022YFB4400600","funder_id":"https://openalex.org/F4320335777","funder_display_name":"National Key Research and Development Program of China"},{"id":"https://openalex.org/G6462694272","display_name":null,"funder_award_id":"2023A007","funder_id":"https://openalex.org/F4320327780","funder_display_name":"Key Research and Development Program of Jiangxi Province"},{"id":"https://openalex.org/G7961932763","display_name":null,"funder_award_id":"KYCX24_0149","funder_id":"https://openalex.org/F4320327780","funder_display_name":"Key Research and Development Program of Jiangxi Province"}],"funders":[{"id":"https://openalex.org/F4320327780","display_name":"Key Research and Development Program of Jiangxi Province","ror":null},{"id":"https://openalex.org/F4320335777","display_name":"National Key Research and Development Program of China","ror":null}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":43,"referenced_works":["https://openalex.org/W2108598243","https://openalex.org/W2340897893","https://openalex.org/W2896457183","https://openalex.org/W2962298324","https://openalex.org/W2963122961","https://openalex.org/W2963163009","https://openalex.org/W2982041622","https://openalex.org/W2982083293","https://openalex.org/W2998218113","https://openalex.org/W3008408165","https://openalex.org/W3129147932","https://openalex.org/W3138516171","https://openalex.org/W3189877953","https://openalex.org/W3206453033","https://openalex.org/W4214588794","https://openalex.org/W4214686755","https://openalex.org/W4285601701","https://openalex.org/W4308479898","https://openalex.org/W4321637298","https://openalex.org/W4385245566","https://openalex.org/W4386076285","https://openalex.org/W4387789586","https://openalex.org/W4390874124","https://openalex.org/W4400524700","https://openalex.org/W6760069825","https://openalex.org/W6767627205","https://openalex.org/W6776767859","https://openalex.org/W6779405248","https://openalex.org/W6780343938","https://openalex.org/W6786271619","https://openalex.org/W6788001715","https://openalex.org/W6788135285","https://openalex.org/W6789240164","https://openalex.org/W6797854001","https://openalex.org/W6802648153","https://openalex.org/W6811272755","https://openalex.org/W6839366308","https://openalex.org/W6845380128","https://openalex.org/W6846125558","https://openalex.org/W6847478871","https://openalex.org/W6848185429","https://openalex.org/W6850162387","https://openalex.org/W6852337893"],"related_works":["https://openalex.org/W2748952813","https://openalex.org/W3107204728","https://openalex.org/W4287591324","https://openalex.org/W3108503355","https://openalex.org/W4226420367","https://openalex.org/W2962876041","https://openalex.org/W3090555870","https://openalex.org/W3022820045","https://openalex.org/W2801655600","https://openalex.org/W2864363823"],"abstract_inverted_index":{"Motivated":[0],"by":[1],"the":[2,8,57,99,110,128,139,157,166,183],"huge":[3,33],"success":[4],"of":[5,10,59,143,171,185],"Transformers":[6,16],"in":[7,26,75],"field":[9],"natural":[11],"language":[12],"processing":[13],"(NLP),":[14],"Vision":[15],"(ViTs)":[17],"have":[18],"been":[19],"rapidly":[20],"developed":[21],"and":[22,36,61,101,122],"achieved":[23],"remarkable":[24],"performance":[25],"various":[27],"computer":[28],"vision":[29],"tasks.":[30],"However,":[31],"their":[32],"model":[34,48],"sizes":[35],"intensive":[37],"computations":[38],"hinder":[39],"ViTs\u2019":[40],"deployment":[41],"on":[42],"embedded":[43],"devices,":[44],"calling":[45],"for":[46,103],"effective":[47],"compression":[49],"methods,":[50],"such":[51],"as":[52,209],"quantization.":[53],"Unfortunately,":[54],"due":[55],"to":[56,70,88,151,165],"existence":[58],"hardware-unfriendly":[60],"quantization-sensitive":[62],"non-linear":[63],"operations,":[64],"particularly":[65],"Softmax,":[66],"it":[67],"is":[68],"non-trivial":[69],"completely":[71],"quantize":[72],"all":[73],"operations":[74],"ViTs,":[76,93,105,173],"yielding":[77],"either":[78],"significant":[79],"accuracy":[80,204],"drops":[81],"or":[82],"non-negligible":[83],"hardware":[84,158,176],"costs.":[85],"In":[86],"response":[87],"challenges":[89],"associated":[90],"with":[91,118],"standard":[92],"we":[94,131,160,190],"focus":[95],"our":[96,186],"attention":[97,117],"towards":[98],"quantization":[100,136,153],"acceleration":[102],"efficient":[104,145,172],"which":[106],"not":[107],"only":[108],"eliminate":[109],"troublesome":[111],"Softmax":[112],"but":[113],"also":[114],"integrate":[115],"linear":[116],"low":[119],"computational":[120],"complexity,":[121],"propose":[123],"Trio-ViT":[124,187],"accordingly.":[125],"Specifically,":[126],"at":[127,156],"algorithm":[129],"level,":[130,159],"develop":[132],"a":[133],"tailored":[134],"post-training":[135],"engine":[137],"taking":[138],"unique":[140],"activation":[141],"distributions":[142],"Softmax-free":[144],"ViTs":[146],"into":[147],"full":[148],"consideration,":[149],"aiming":[150],"boost":[152],"accuracy.":[154],"Furthermore,":[155],"build":[161],"an":[162],"accelerator":[163],"dedicated":[164],"specific":[167],"Convolution-Transformer":[168],"hybrid":[169],"architecture":[170],"thereby":[174],"enhancing":[175],"efficiency.":[177,219],"Extensive":[178],"experimental":[179],"results":[180],"consistently":[181],"prove":[182],"effectiveness":[184],"framework.":[188],"Particularly,":[189],"can":[191],"gain":[192],"up":[193],"to$\\uparrow":[194],"{3.6}\\times":[195],"$,$\\uparrow":[196,213],"{5.0}\\times":[197],"$,":[198,215],"and$\\uparrow":[199,216],"{7.3}\\times":[200],"$FPS":[201],"under":[202],"comparable":[203],"over":[205],"state-of-the-art":[206],"ViT":[207],"accelerators,":[208],"well":[210],"as$\\uparrow":[211],"{6.0}\\times":[212],"{1.5}\\times":[214],"{2.1}\\times":[217],"$DSP":[218],"Codes":[220],"are":[221],"available":[222],"athttps://github.com/shihuihong214/Trio-ViT.":[223]},"counts_by_year":[{"year":2026,"cited_by_count":4},{"year":2025,"cited_by_count":7}],"updated_date":"2026-08-30T07:30:35.949966","created_date":"2025-10-10T00:00:00"}