{"id":"https://openalex.org/W7160897362","doi":"https://doi.org/10.48550/arxiv.2605.09430","title":"FlashAR: Efficient Post-Training Acceleration for Autoregressive Image Generation","display_name":"FlashAR: Efficient Post-Training Acceleration for Autoregressive Image Generation","publication_year":2026,"publication_date":"2026-05-10","ids":{"openalex":"https://openalex.org/W7160897362","doi":"https://doi.org/10.48550/arxiv.2605.09430"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.09430","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.09430","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.09430","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135981665","display_name":"Junkang Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Junkang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135923338","display_name":"Yefei He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Yefei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135976958","display_name":"Feng Chen","orcid":"https://orcid.org/0000-0003-4432-6005"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Feng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135970076","display_name":"Weijie Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Weijie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135916112","display_name":"Bohan Zhuang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhuang, Bohan","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":null,"has_fulltext":false,"cited_by_count":0,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.7246999740600586,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.7246999740600586,"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/T11019","display_name":"Image Enhancement Techniques","score":0.05700000002980232,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.045499999076128006,"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/autoregressive-model","display_name":"Autoregressive model","score":0.7985000014305115},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.5812000036239624},{"id":"https://openalex.org/keywords/generator","display_name":"Generator (circuit theory)","score":0.5638999938964844},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.5383999943733215},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.4832000136375427},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.48089998960494995},{"id":"https://openalex.org/keywords/head","display_name":"Head (geology)","score":0.47620001435279846}],"concepts":[{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.7985000014305115},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.777999997138977},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.5812000036239624},{"id":"https://openalex.org/C2780992000","wikidata":"https://www.wikidata.org/wiki/Q17016113","display_name":"Generator (circuit theory)","level":3,"score":0.5638999938964844},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.5383999943733215},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.4832000136375427},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.48089998960494995},{"id":"https://openalex.org/C2780312720","wikidata":"https://www.wikidata.org/wiki/Q5689100","display_name":"Head (geology)","level":2,"score":0.47620001435279846},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.46459999680519104},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.4487999975681305},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.43549999594688416},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.43380001187324524},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.3151000142097473},{"id":"https://openalex.org/C198082294","wikidata":"https://www.wikidata.org/wiki/Q3399648","display_name":"Position (finance)","level":2,"score":0.26669999957084656},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.26330000162124634},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.260699987411499},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.25929999351501465}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.09430","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.09430","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.09430","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.09430","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large-scale":[0],"autoregressive":[1,72,207],"models":[2],"have":[3],"demonstrated":[4],"remarkable":[5],"capabilities":[6],"in":[7],"image":[8,238],"generation.":[9],"However,":[10],"their":[11],"sequential":[12],"raster-scan":[13,71],"decoding":[14,219],"relies":[15],"on":[16,80,223],"strictly":[17],"next-token":[18,82],"prediction,":[19],"making":[20],"inference":[21],"prohibitively":[22],"expensive.":[23],"Existing":[24],"acceleration":[25],"methods":[26],"typically":[27],"either":[28],"introduce":[29,59,121],"entirely":[30],"new":[31,218],"generation":[32,43,239],"paradigms":[33],"that":[34,66,88,228],"necessitate":[35],"costly":[36],"pre-training":[37],"from":[38,139,204],"scratch,":[39],"or":[40,51],"enable":[41],"parallel":[42,77],"at":[44,182],"the":[45,95,109,136,145,149,179,196,205,217,248],"expense":[46],"of":[47,247],"a":[48,61,69,75,114,122,172,192,233,241],"training-inference":[49],"gap":[50],"altered":[52],"prediction":[53,119],"objectives.":[54],"In":[55],"this":[56],"paper,":[57],"we":[58,107,134,170,190],"FlashAR,":[60],"lightweight":[62,124,242],"post-training":[63,243],"adaptation":[64,90,188,194,203],"framework":[65],"efficiently":[67],"adapts":[68],"pre-trained":[70,96,206],"model":[73,208],"into":[74],"highly":[76],"generator":[78],"based":[79],"two-way":[81],"prediction.":[83,129],"Our":[84],"key":[85],"insight":[86],"is":[87,199],"effective":[89],"should":[91],"minimize":[92],"modifications":[93],"to":[94,101,176,214,216,232],"model's":[97],"original":[98,110,249],"training":[99,250],"objective":[100],"preserve":[102],"its":[103],"learned":[104],"prior.":[105],"Accordingly,":[106],"retain":[108],"AR":[111],"head":[112,116,126,138,152,198],"as":[113],"horizontal":[115,151,156],"for":[117,127,236],"row-wise":[118],"and":[120,157,225],"complementary,":[123],"vertical":[125,137,158,197],"column-wise":[128],"To":[130,185],"facilitate":[131],"efficient":[132],"adaptation,":[133],"branch":[135],"an":[140],"intermediate":[141],"layer":[142],"rather":[143],"than":[144],"final":[146],"layer,":[147],"bypassing":[148],"inherent":[150],"bias.":[153],"Moreover,":[154],"since":[155],"predictions":[159,181],"capture":[160],"complementary":[161],"dependencies":[162],"whose":[163],"relative":[164],"importance":[165],"varies":[166],"across":[167],"target":[168],"positions,":[169],"employ":[171],"learnable":[173],"fusion":[174],"gate":[175],"dynamically":[177],"combine":[178],"two":[180],"each":[183],"position.":[184],"further":[186],"reduce":[187],"cost,":[189],"propose":[191],"two-stage":[193],"pipeline:":[195],"first":[200],"initialized":[201],"through":[202,240],"before":[209],"jointly":[210],"fine-tuned":[211],"with":[212,244],"backbone":[213],"adapt":[215],"paradigm.":[220],"Extensive":[221],"experiments":[222],"LlamaGen":[224],"Emu3.5":[226],"show":[227],"FlashAR":[229],"achieves":[230],"up":[231],"22.9x":[234],"speedup":[235],"512x512":[237],"merely":[245],"0.05%":[246],"data.":[251]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-13T00:00:00"}