{"id":"https://openalex.org/W4306886919","doi":"https://doi.org/10.48550/arxiv.2210.09461","title":"Token Merging: Your ViT But Faster","display_name":"Token Merging: Your ViT But Faster","publication_year":2022,"publication_date":"2022-10-17","ids":{"openalex":"https://openalex.org/W4306886919","doi":"https://doi.org/10.48550/arxiv.2210.09461"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2210.09461","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2210.09461","pdf_url":"https://arxiv.org/pdf/2210.09461","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2210.09461","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5063326998","display_name":"Daniel Bolya","orcid":"https://orcid.org/0000-0003-0223-3599","observed_orcids":["https://orcid.org/0000-0003-0223-3599"]},"institutions":[{"id":"https://openalex.org/I130701444","display_name":"Georgia Institute of Technology","ror":"https://ror.org/01zkghx44","country_code":"US","type":"education","lineage":["https://openalex.org/I130701444"]},{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Bolya, Daniel","raw_affiliation_strings":["Georgia Tech","Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Georgia Tech","institution_ids":["https://openalex.org/I130701444"]},{"raw_affiliation_string":"Meta AI","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006183921","display_name":"Cheng-Yang Fu","orcid":"https://orcid.org/0000-0002-1377-3278","observed_orcids":["https://orcid.org/0000-0002-1377-3278"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Fu, Cheng-Yang","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080137972","display_name":"Xiaoliang Dai","orcid":"https://orcid.org/0000-0003-3098-2714","observed_orcids":["https://orcid.org/0000-0003-3098-2714"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dai, Xiaoliang","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057613852","display_name":"Peizhao Zhang","orcid":"https://orcid.org/0000-0001-7128-191X","observed_orcids":["https://orcid.org/0000-0001-7128-191X"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhang, Peizhao","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5036069974","display_name":"Christoph Feichtenhofer","orcid":"https://orcid.org/0000-0001-9756-7238","observed_orcids":["https://orcid.org/0000-0001-9756-7238"]},"institutions":[{"id":"https://openalex.org/I4210114444","display_name":"Meta (United States)","ror":"https://ror.org/01zbnvs85","country_code":"US","type":"company","lineage":["https://openalex.org/I4210114444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Feichtenhofer, Christoph","raw_affiliation_strings":["Meta AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Meta AI","institution_ids":["https://openalex.org/I4210114444"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5038068459","display_name":"Judy Hoffman","orcid":"https://orcid.org/0000-0003-1971-1606","observed_orcids":["https://orcid.org/0000-0003-1971-1606"]},"institutions":[{"id":"https://openalex.org/I130701444","display_name":"Georgia Institute of Technology","ror":"https://ror.org/01zkghx44","country_code":"US","type":"education","lineage":["https://openalex.org/I130701444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hoffman, Judy","raw_affiliation_strings":["Georgia Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Georgia Tech","institution_ids":["https://openalex.org/I130701444"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":684,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11605","display_name":"Visual Attention and Saliency Detection","score":0.0,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/token-merging","display_name":"token merging","score":0.996999979019165},{"id":"https://openalex.org/keywords/vision-transformer","display_name":"vision transformer","score":0.8209999799728394},{"id":"https://openalex.org/keywords/training-acceleration","display_name":"training acceleration","score":0.593999981880188},{"id":"https://openalex.org/keywords/audio-representation-learning","display_name":"audio representation learning","score":0.42899999022483826},{"id":"https://openalex.org/keywords/video-representation-learning","display_name":"video representation learning","score":0.843999981880188}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6916211843490601},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6716087460517883},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.6463073492050171},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5441920757293701},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.4285024404525757},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3468558192253113},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3247019648551941},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.2615150809288025},{"id":"https://openalex.org/C555944384","wikidata":"https://www.wikidata.org/wiki/Q249","display_name":"Wireless","level":2,"score":0.22545126080513},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.14066088199615479},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.13132894039154053},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2210.09461","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2210.09461","pdf_url":"https://arxiv.org/pdf/2210.09461","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2210.09461","doi":"https://doi.org/10.48550/arxiv.2210.09461","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2210.09461","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2210.09461","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2210.09461","pdf_url":"https://arxiv.org/pdf/2210.09461","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4306886919.pdf","grobid_xml":"https://content.openalex.org/works/W4306886919.grobid-xml"},"referenced_works_count":11,"referenced_works":["https://openalex.org/W2619947201","https://openalex.org/W3033529678","https://openalex.org/W3091156754","https://openalex.org/W3174708387","https://openalex.org/W3180037928","https://openalex.org/W3205814820","https://openalex.org/W4221144813","https://openalex.org/W4226050675","https://openalex.org/W4320036893","https://openalex.org/W4365446402","https://openalex.org/W7133227460"],"related_works":["https://openalex.org/W4414198544","https://openalex.org/W4323556158","https://openalex.org/W4384159680","https://openalex.org/W4415708394","https://openalex.org/W7162635299","https://openalex.org/W7154626294","https://openalex.org/W7202142785","https://openalex.org/W4388720130","https://openalex.org/W4410476912","https://openalex.org/W4415797474"],"abstract_inverted_index":{"We":[0],"introduce":[1],"Token":[2],"Merging":[3],"(ToMe),":[4],"a":[5,26,29,73,119],"simple":[6],"method":[7],"to":[8,18,94,109],"increase":[9],"the":[10,49,65,111],"throughput":[11,50,66,112],"of":[12,51,67,113,138],"existing":[13],"ViT":[14],"models":[15,60],"without":[16],"needing":[17],"train.":[19],"ToMe":[20,46,80,103,127],"gradually":[21],"combines":[22],"similar":[23],"tokens":[24],"in":[25,77,89],"transformer":[27],"using":[28],"general":[30],"and":[31,56,63,143,152],"light-weight":[32],"matching":[33],"algorithm":[34],"that":[35,126],"is":[36],"as":[37,39],"fast":[38],"pruning":[40],"while":[41],"being":[42],"more":[43],"accurate.":[44],"Off-the-shelf,":[45],"can":[47,81],"2x":[48,95,110],"state-of-the-art":[52,148],"ViT-L":[53,68],"@":[54,58],"512":[55],"ViT-H":[57],"518":[59],"on":[61,69,99,115,149],"images":[62],"2.2x":[64],"video":[70],"with":[71,102,147],"only":[72,118],"0.2-0.3%":[74],"accuracy":[75,106,142],"drop":[76],"each":[78],"case.":[79],"also":[82],"easily":[83],"be":[84],"applied":[85],"during":[86],"training,":[87],"improving":[88],"practice":[90],"training":[91],"speed":[92,144],"up":[93],"for":[96,117],"MAE":[97],"fine-tuning":[98],"video.":[100,139],"Training":[101],"further":[104],"minimizes":[105],"drop,":[107],"leading":[108],"ViT-B":[114],"audio":[116],"0.4%":[120],"mAP":[121],"drop.":[122],"Qualitatively,":[123],"we":[124],"find":[125],"merges":[128],"object":[129],"parts":[130],"into":[131],"one":[132],"token,":[133],"even":[134],"over":[135],"multiple":[136],"frames":[137],"Overall,":[140],"ToMe's":[141],"are":[145],"competitive":[146],"images,":[150],"video,":[151],"audio.":[153]},"counts_by_year":[{"year":2026,"cited_by_count":297},{"year":2025,"cited_by_count":213},{"year":2024,"cited_by_count":134},{"year":2023,"cited_by_count":39},{"year":2022,"cited_by_count":1}],"updated_date":"2026-11-12T12:28:57.059296","created_date":"2022-10-21T00:00:00"}