{"id":"https://openalex.org/W4386075861","doi":"https://doi.org/10.1109/cvpr52729.2023.00359","title":"TinyMIM: An Empirical Study of Distilling MIM Pre-trained Models","display_name":"TinyMIM: An Empirical Study of Distilling MIM Pre-trained Models","publication_year":2023,"publication_date":"2023-06-01","ids":{"openalex":"https://openalex.org/W4386075861","doi":"https://doi.org/10.1109/cvpr52729.2023.00359"},"language":"en","primary_location":{"id":"doi:10.1109/cvpr52729.2023.00359","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvpr52729.2023.00359","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","raw_type":"proceedings-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":"https://openalex.org/A5048891976","display_name":"Sucheng Ren","orcid":"https://orcid.org/0000-0003-4730-8435"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Sucheng Ren","raw_affiliation_strings":["Microsoft Research Asia"],"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090973869","display_name":"Fangyun Wei","orcid":"https://orcid.org/0000-0001-8784-4916"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fangyun Wei","raw_affiliation_strings":["Microsoft Research Asia"],"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100459168","display_name":"Zheng Zhang","orcid":"https://orcid.org/0000-0003-1470-6998"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zheng Zhang","raw_affiliation_strings":["Microsoft Research Asia"],"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101976576","display_name":"Han Hu","orcid":"https://orcid.org/0000-0001-5104-6146"},"institutions":[{"id":"https://openalex.org/I4210113369","display_name":"Microsoft Research Asia (China)","ror":"https://ror.org/0300m5276","country_code":"CN","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210113369"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Han Hu","raw_affiliation_strings":["Microsoft Research Asia"],"affiliations":[{"raw_affiliation_string":"Microsoft Research Asia","institution_ids":["https://openalex.org/I4210113369"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5048891976"],"corresponding_institution_ids":["https://openalex.org/I4210113369"],"apc_list":null,"apc_paid":null,"fwci":5.3443,"has_fulltext":false,"cited_by_count":44,"citation_normalized_percentile":{"value":0.96954136,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"3687","last_page":"3697"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":1.0,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":1.0,"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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9991000294685364,"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/T10627","display_name":"Advanced Image and Video Retrieval Techniques","score":0.9988999962806702,"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/computer-science","display_name":"Computer science","score":0.7638113498687744},{"id":"https://openalex.org/keywords/regularization","display_name":"Regularization (linguistics)","score":0.65630704164505},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.621437668800354},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.6119575500488281},{"id":"https://openalex.org/keywords/cls-upper-limits","display_name":"CLs upper limits","score":0.5663970708847046},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.5661239624023438},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.5120342373847961},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.5050123333930969},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4927861988544464},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.46547549962997437},{"id":"https://openalex.org/keywords/segmentation","display_name":"Segmentation","score":0.4513987898826599},{"id":"https://openalex.org/keywords/contextual-image-classification","display_name":"Contextual image classification","score":0.4344256818294525},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.3427085876464844},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.26640626788139343},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.20866984128952026},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.07476046681404114}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7638113498687744},{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.65630704164505},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.621437668800354},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.6119575500488281},{"id":"https://openalex.org/C190729725","wikidata":"https://www.wikidata.org/wiki/Q5012817","display_name":"CLs upper limits","level":2,"score":0.5663970708847046},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.5661239624023438},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.5120342373847961},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5050123333930969},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4927861988544464},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.46547549962997437},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.4513987898826599},{"id":"https://openalex.org/C75294576","wikidata":"https://www.wikidata.org/wiki/Q5165192","display_name":"Contextual image classification","level":3,"score":0.4344256818294525},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3427085876464844},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.26640626788139343},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.20866984128952026},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07476046681404114},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C165801399","wikidata":"https://www.wikidata.org/wiki/Q25428","display_name":"Voltage","level":2,"score":0.0},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","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/C119767625","wikidata":"https://www.wikidata.org/wiki/Q618211","display_name":"Optometry","level":1,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/cvpr52729.2023.00359","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvpr52729.2023.00359","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":88,"referenced_works":["https://openalex.org/W1690739335","https://openalex.org/W1821462560","https://openalex.org/W2739879705","https://openalex.org/W2743289088","https://openalex.org/W2789007410","https://openalex.org/W2803023299","https://openalex.org/W2886756692","https://openalex.org/W2896457183","https://openalex.org/W2903396356","https://openalex.org/W2955425717","https://openalex.org/W2963534679","https://openalex.org/W2964111476","https://openalex.org/W2982083293","https://openalex.org/W2982157312","https://openalex.org/W3004127093","https://openalex.org/W3034368386","https://openalex.org/W3034619766","https://openalex.org/W3121052760","https://openalex.org/W3129576130","https://openalex.org/W3131500599","https://openalex.org/W3138154797","https://openalex.org/W3138516171","https://openalex.org/W3145450063","https://openalex.org/W3159481202","https://openalex.org/W3166396011","https://openalex.org/W3167917117","https://openalex.org/W3170863103","https://openalex.org/W3170874841","https://openalex.org/W3190492058","https://openalex.org/W3213836217","https://openalex.org/W3215434919","https://openalex.org/W4214588794","https://openalex.org/W4214653736","https://openalex.org/W4221144677","https://openalex.org/W4221145109","https://openalex.org/W4281702809","https://openalex.org/W4281756776","https://openalex.org/W4283807635","https://openalex.org/W4286910290","https://openalex.org/W4292112947","https://openalex.org/W4297775537","https://openalex.org/W4300973271","https://openalex.org/W4300978573","https://openalex.org/W4301914798","https://openalex.org/W4307929719","https://openalex.org/W4312262772","https://openalex.org/W4312273141","https://openalex.org/W4312312750","https://openalex.org/W4312349930","https://openalex.org/W4312599584","https://openalex.org/W4312804044","https://openalex.org/W4313156423","https://openalex.org/W4313160444","https://openalex.org/W4313170858","https://openalex.org/W4386065386","https://openalex.org/W4386072059","https://openalex.org/W4386076385","https://openalex.org/W4386221015","https://openalex.org/W6637551013","https://openalex.org/W6638523607","https://openalex.org/W6737664043","https://openalex.org/W6748082341","https://openalex.org/W6751751081","https://openalex.org/W6755207826","https://openalex.org/W6756633790","https://openalex.org/W6762718338","https://openalex.org/W6788135285","https://openalex.org/W6790978476","https://openalex.org/W6791353385","https://openalex.org/W6796761347","https://openalex.org/W6802648153","https://openalex.org/W6804160461","https://openalex.org/W6804229649","https://openalex.org/W6810007534","https://openalex.org/W6810107437","https://openalex.org/W6810423238","https://openalex.org/W6810744310","https://openalex.org/W6837873513","https://openalex.org/W6838376339","https://openalex.org/W6838637664","https://openalex.org/W6838673894","https://openalex.org/W6838706884","https://openalex.org/W6839178539","https://openalex.org/W6841885045","https://openalex.org/W6844617645","https://openalex.org/W6844713038","https://openalex.org/W6845452229","https://openalex.org/W6846302668"],"related_works":["https://openalex.org/W2532775738","https://openalex.org/W2546942002","https://openalex.org/W2970216048","https://openalex.org/W2382607599","https://openalex.org/W2742991909","https://openalex.org/W2509146328","https://openalex.org/W133358225","https://openalex.org/W2508908072","https://openalex.org/W1577137544","https://openalex.org/W2996038082"],"abstract_inverted_index":{"Masked":[0],"image":[1,180],"modeling":[2],"(MIM)":[3],"performs":[4],"strongly":[5],"in":[6,54,156,226],"pretraining":[7],"large":[8,42],"vision":[9,189,208],"Transformers":[10],"(ViTs).":[11],"However,":[12],"small":[13,188,207],"models":[14,45,190],"that":[15,96,108,211],"are":[16],"critical":[17],"for":[18,187,205],"real-world":[19],"applications":[20],"can-not":[21],"or":[22],"only":[23],"marginally":[24],"benefit":[25],"from":[26],"this":[27,31],"pre-training":[28,131],"approach.":[29],"In":[30],"paper,":[32],"we":[33,121],"explore":[34],"distillation":[35,56],"techniques":[36],"to":[37,46],"transfer":[38],"the":[39,55,88,98,102,105,110,128,137,165,192],"success":[40],"of":[41,87,104,109,150,171,191],"MIM-based":[43],"pre-trained":[44],"smaller":[47],"ones.":[48],"We":[49],"systematically":[50],"study":[51],"different":[52],"options":[53],"framework,":[57],"including":[58],"distilling":[59],"targets,":[60],"losses,":[61],"input,":[62],"network":[63,90],"regularization,":[64],"sequential":[65],"distillation,":[66],"etc,":[67],"revealing":[68],"that:":[69],"1)":[70],"Distilling":[71],"token":[72],"relations":[73],"is":[74,115,161,231],"more":[75],"effective":[76],"than":[77,95,164,219],"CLS":[78],"token-":[79],"and":[80,140,195],"feature-based":[81],"distillation;":[82],"2)":[83],"An":[84],"intermediate":[85],"layer":[86,100],"teacher":[89],"as":[91,225],"target":[92],"perform":[93],"better":[94,215],"using":[97,135],"last":[99],"when":[101],"depth":[103],"student":[106],"mismatches":[107],"teacher;":[111],"3)":[112],"Weak":[113],"regularization":[114],"preferred;":[116],"etc.":[117],"With":[118],"these":[119],"findings,":[120],"achieve":[122],"significant":[123],"fine-tuning":[124],"accuracy":[125,177],"improvements":[126],"over":[127],"scratch":[129],"MIM":[130],"on":[132,178],"ImageNet-1K":[133,179],"classification,":[134,181],"all":[136],"ViT-Tiny,":[138],"ViT-Small,":[139],"ViT-base":[141],"models,":[142,210],"with":[143],"+4.2%/+2.4%/+1.4%":[144],"gains,":[145],"respectively.":[146],"Our":[147,168],"TinyMIM":[148,169],"model":[149,170],"base":[151],"size":[152,173,194],"achieves":[153,174],"52.2":[154],"mIoU":[155],"AE20K":[157],"semantic":[158],"segmentation,":[159],"which":[160,182],"+4.1":[162],"higher":[163],"MAE":[166],"baseline.":[167],"tiny":[172],"79.6%":[175],"top-1":[176],"sets":[183],"a":[184],"new":[185],"record":[186],"same":[193],"computation":[196],"budget.":[197],"This":[198],"strong":[199],"performance":[200],"suggests":[201],"an":[202],"alternative":[203],"way":[204],"developing":[206],"Transformer":[209],"is,":[212],"by":[213],"exploring":[214],"training":[216],"methods":[217],"rather":[218],"introducing":[220],"inductive":[221],"biases":[222],"into":[223],"architectures":[224],"most":[227],"previous":[228],"works.":[229],"Code":[230],"available":[232],"at":[233],"https://github.com/OliverRensu/TinyMIM.":[234]},"counts_by_year":[{"year":2026,"cited_by_count":4},{"year":2025,"cited_by_count":19},{"year":2024,"cited_by_count":16},{"year":2023,"cited_by_count":5}],"updated_date":"2026-03-20T23:20:44.827607","created_date":"2025-10-10T00:00:00"}