{"id":"https://openalex.org/W4413144648","doi":"https://doi.org/10.1109/cvpr52734.2025.00910","title":"Token Cropr: Faster ViTs for Quite a Few Tasks","display_name":"Token Cropr: Faster ViTs for Quite a Few Tasks","publication_year":2025,"publication_date":"2025-06-10","ids":{"openalex":"https://openalex.org/W4413144648","doi":"https://doi.org/10.1109/cvpr52734.2025.00910"},"language":"en","primary_location":{"id":"doi:10.1109/cvpr52734.2025.00910","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvpr52734.2025.00910","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5012321770","display_name":"Benjamin Bergner","orcid":"https://orcid.org/0000-0002-0572-840X"},"institutions":[{"id":"https://openalex.org/I143288331","display_name":"Hasso Plattner Institute","ror":"https://ror.org/058rn5r42","country_code":"DE","type":"facility","lineage":["https://openalex.org/I143288331","https://openalex.org/I176453806"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Benjamin Bergner","raw_affiliation_strings":["University of Potsdam,Hasso Plattner Institute for Digital Engineering"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Potsdam,Hasso Plattner Institute for Digital Engineering","institution_ids":["https://openalex.org/I143288331"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074204533","display_name":"Christoph Lippert","orcid":"https://orcid.org/0000-0001-6363-2556"},"institutions":[{"id":"https://openalex.org/I143288331","display_name":"Hasso Plattner Institute","ror":"https://ror.org/058rn5r42","country_code":"DE","type":"facility","lineage":["https://openalex.org/I143288331","https://openalex.org/I176453806"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Christoph Lippert","raw_affiliation_strings":["University of Potsdam,Hasso Plattner Institute for Digital Engineering"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Potsdam,Hasso Plattner Institute for Digital Engineering","institution_ids":["https://openalex.org/I143288331"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5051934274","display_name":"Aravindh Mahendran","orcid":"https://orcid.org/0000-0002-2650-9871"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]},{"id":"https://openalex.org/I4210090411","display_name":"Google DeepMind (United Kingdom)","ror":"https://ror.org/00971b260","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210090411","https://openalex.org/I4210128969"]}],"countries":["GB","US"],"is_corresponding":false,"raw_author_name":"Aravindh Mahendran","raw_affiliation_strings":["Google DeepMind"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google DeepMind","institution_ids":["https://openalex.org/I1291425158","https://openalex.org/I4210090411"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"9740","last_page":"9750"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.7293999791145325,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.7293999791145325,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.7143999934196472,"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/T10789","display_name":"Interactive and Immersive Displays","score":0.6974999904632568,"subfield":{"id":"https://openalex.org/subfields/1709","display_name":"Human-Computer Interaction"},"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":0.7991243600845337},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.6324151754379272},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.4747981131076813},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.26691728830337524}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7991243600845337},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6324151754379272},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.4747981131076813},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.26691728830337524}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/cvpr52734.2025.00910","is_oa":false,"landing_page_url":"https://doi.org/10.1109/cvpr52734.2025.00910","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 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":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W4388335561","https://openalex.org/W2970530566","https://openalex.org/W4288261899","https://openalex.org/W4307309205","https://openalex.org/W2967478618","https://openalex.org/W4385009901","https://openalex.org/W4385572700"],"abstract_inverted_index":{"The":[0],"adoption":[1],"of":[2,34,98,119,153],"Vision":[3],"Transformers":[4],"(ViTs)":[5],"in":[6,11,126],"resource-constrained":[7],"applications":[8],"necessitates":[9],"improvements":[10],"inference":[12],"throughput.":[13],"To":[14],"this":[15,62],"end":[16],"several":[17],"token":[18,45,67],"pruning":[19],"and":[20,54,113,116],"merging":[21],"approaches":[22],"have":[23],"been":[24],"proposed":[25],"that":[26,48,69,74,97],"improve":[27],"efficiency":[28],"by":[29],"successively":[30],"reducing":[31],"the":[32,133,145],"number":[33],"tokens.":[35],"However,":[36],"it":[37],"remains":[38],"an":[39],"open":[40],"problem":[41],"to":[42,57,76,93,96,144],"design":[43],"a":[44,66,99,129,140,149],"reduction":[46],"method":[47,105],"is":[49,55],"fast,":[50],"maintains":[51],"high":[52],"performance,":[53],"applicable":[56],"various":[58],"vision":[59],"tasks.":[60],"In":[61],"work,":[63],"we":[64,138],"present":[65],"pruner":[68],"uses":[70],"auxiliary":[71,85],"prediction":[72],"heads":[73,86],"learn":[75],"select":[77],"tokens":[78],"end-to-end":[79],"based":[80],"on":[81,106,132],"task":[82],"relevance.":[83],"These":[84],"can":[87],"be":[88],"removed":[89],"after":[90],"training,":[91],"leading":[92],"throughput":[94],"close":[95],"random":[100],"pruner.":[101],"We":[102],"evaluate":[103],"our":[104],"image":[107],"classification,":[108],"semantic":[109,135],"segmentation,":[110,115],"object":[111],"detection,":[112],"instance":[114],"show":[117],"speedups":[118],"1.5":[120],"\u2013":[121],"4x":[122],"with":[123,148],"small":[124],"drops":[125],"performance.":[127],"As":[128],"best":[130],"case,":[131],"ADE20k":[134],"segmentation":[136],"benchmark,":[137],"observe":[139],"2x":[141],"speedup":[142],"relative":[143],"no-pruning":[146],"baseline,":[147],"negligible":[150],"performance":[151],"penalty":[152],"0.1":[154],"median":[155],"mIoU":[156],"across":[157],"5":[158],"seeds.":[159]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-25T09:21:30.201066","created_date":"2025-10-10T00:00:00"}