{"id":"https://openalex.org/W4376983091","doi":null,"title":"Training data-efficient image transformers & distillation through attention","display_name":"Training data-efficient image transformers & distillation through attention","publication_year":2021,"publication_date":"2021-07-19","ids":{"openalex":"https://openalex.org/W4376983091"},"language":"en","primary_location":{"id":"pmh:oai:HAL:hal-03997937v1","doi":null,"is_oa":false,"landing_page_url":"https://hal.science/hal-03997937","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I4405252976","host_organization_name":"Centre pour la Communication Scientifique Directe","host_organization_lineage":["https://openalex.org/I4405252976"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ICML, Jul 2021, visio, France","raw_type":"info:eu-repo/semantics/conferenceObject"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":[],"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":"Hugo, Touvron","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hugo, Touvron","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108118084","display_name":"Matthieu Cord","orcid":"https://orcid.org/0000-0002-0627-5844","observed_orcids":["https://orcid.org/0000-0002-0627-5844"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cord, Matthieu","raw_affiliation_strings":["Machine Learning and Information Access"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Machine Learning and Information Access","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053822883","display_name":"Matthijs Douze","orcid":"https://orcid.org/0009-0009-6754-7821","observed_orcids":["https://orcid.org/0009-0009-6754-7821"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Matthijs, Douze","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088298158","display_name":"Massa Francisco","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Francisco, Massa","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Alexandre, Sablayrolles","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Alexandre, Sablayrolles","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5111802678","display_name":"Herv\u00e9 Je\u01f5ou","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Herve, Jegou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.0927,"has_fulltext":false,"cited_by_count":36,"citation_normalized_percentile":{"value":0.91693949,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":87,"max":99},"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":0.7333999872207642,"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":0.7333999872207642,"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":0.26100000739097595,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.0017000000225380063,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/image-transformer","display_name":"image transformer","score":0.878000020980835},{"id":"https://openalex.org/keywords/data-efficient-training","display_name":"data-efficient training","score":0.7289999723434448},{"id":"https://openalex.org/keywords/knowledge-distillation","display_name":"knowledge distillation","score":0.9300000071525574},{"id":"https://openalex.org/keywords/attention-mechanism","display_name":"attention mechanism","score":0.7369999885559082},{"id":"https://openalex.org/keywords/vision-transformer","display_name":"vision transformer","score":0.47099998593330383}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6670201420783997},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.6224915981292725},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5954132676124573},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4669400751590729},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.42537933588027954},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4243120551109314},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.41299062967300415},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.1622670590877533},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.13580095767974854},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.07709154486656189},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C178790620","wikidata":"https://www.wikidata.org/wiki/Q11351","display_name":"Organic chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:HAL:hal-03997937v1","doi":null,"is_oa":false,"landing_page_url":"https://hal.science/hal-03997937","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I4405252976","host_organization_name":"Centre pour la Communication Scientifique Directe","host_organization_lineage":["https://openalex.org/I4405252976"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ICML, Jul 2021, visio, France","raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4310614830","https://openalex.org/W4312677967","https://openalex.org/W4312671791","https://openalex.org/W3119786062","https://openalex.org/W4377692680","https://openalex.org/W4402915866","https://openalex.org/W4376983087","https://openalex.org/W4313303591","https://openalex.org/W4281488278","https://openalex.org/W7133185110"],"abstract_inverted_index":{"Recently,":[0],"neural":[1],"networks":[2],"purely":[3],"based":[4],"on":[5,47,71,88,121],"attention":[6],"were":[7],"shown":[8],"to":[9,84,127],"address":[10],"image":[11,16],"understanding":[12],"tasks":[13],"such":[14],"as":[15],"classification.":[17],"These":[18],"highperforming":[19],"vision":[20,61],"transformers":[21,45,109],"are":[22,110],"pre-trained":[23],"with":[24,73,115],"hundreds":[25],"of":[26,28,68,118],"millions":[27],"images":[29],"using":[30,50],"a":[31,51,80,89,104],"large":[32],"infrastructure,":[33],"thereby":[34],"limiting":[35],"their":[36],"adoption.":[37],"In":[38],"this":[39],"work,":[40],"we":[41],"produce":[42],"competitive":[43,111],"convolutionfree":[44],"trained":[46],"ImageNet":[48,72],"only":[49],"single":[52],"computer":[53],"in":[54],"less":[55],"than":[56],"3":[57],"days.":[58],"Our":[59],"reference":[60],"transformer":[62],"(86M":[63],"parameters)":[64],"achieves":[65],"top-1":[66,113],"accuracy":[67],"83.1%":[69],"(single-crop)":[70],"no":[74],"external":[75],"data.":[76],"We":[77,130],"also":[78],"introduce":[79],"teacher-student":[81],"strategy":[82],"specific":[83],"transformers.":[85],"It":[86],"relies":[87],"distillation":[90],"token":[91],"ensuring":[92],"that":[93],"the":[94,98,116,119],"student":[95],"learns":[96],"from":[97,103],"teacher":[99],"through":[100],"attention,":[101],"typically":[102],"convnet":[105],"teacher.":[106],"The":[107],"learned":[108],"(85.2%":[112],"acc.)":[114],"state":[117],"art":[120],"ImageNet,":[122],"and":[123,135],"similarly":[124],"when":[125],"transferred":[126],"other":[128],"tasks.":[129],"will":[131],"share":[132],"our":[133],"code":[134],"models.":[136]},"counts_by_year":[{"year":2025,"cited_by_count":9},{"year":2024,"cited_by_count":13},{"year":2023,"cited_by_count":11},{"year":2022,"cited_by_count":2},{"year":2021,"cited_by_count":1}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2025-10-10T00:00:00"}