{"id":"https://openalex.org/W4393924573","doi":"https://doi.org/10.48550/arxiv.2404.01367","title":"Bigger is not Always Better: Scaling Properties of Latent Diffusion Models","display_name":"Bigger is not Always Better: Scaling Properties of Latent Diffusion Models","publication_year":2024,"publication_date":"2024-04-01","ids":{"openalex":"https://openalex.org/W4393924573","doi":"https://doi.org/10.48550/arxiv.2404.01367"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2404.01367","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2404.01367","pdf_url":"https://arxiv.org/pdf/2404.01367","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":null,"license_id":null,"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/2404.01367","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5061336891","display_name":"Kangfu Mei","orcid":"https://orcid.org/0000-0001-8949-9597","observed_orcids":["https://orcid.org/0000-0001-8949-9597"]},"institutions":[{"id":"https://openalex.org/I145311948","display_name":"Johns Hopkins University","ror":"https://ror.org/00za53h95","country_code":"US","type":"education","lineage":["https://openalex.org/I145311948"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Mei, Kangfu","raw_affiliation_strings":["Johns Hopkins University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Johns Hopkins University","institution_ids":["https://openalex.org/I145311948"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5015173810","display_name":"Zhengzhong Tu","orcid":"https://orcid.org/0000-0002-7594-2292","observed_orcids":["https://orcid.org/0000-0002-7594-2292"]},"institutions":[{"id":"https://openalex.org/I91045830","display_name":"Texas A&M University","ror":"https://ror.org/01f5ytq51","country_code":"US","type":"education","lineage":["https://openalex.org/I91045830"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tu, Zhengzhong","raw_affiliation_strings":["Texas A&M University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Texas A&M University","institution_ids":["https://openalex.org/I91045830"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084566603","display_name":"Mauricio Delbracio","orcid":"https://orcid.org/0000-0001-7539-2991","observed_orcids":["https://orcid.org/0000-0001-7539-2991"]},"institutions":[{"id":"https://openalex.org/I91045830","display_name":"Texas A&M University","ror":"https://ror.org/01f5ytq51","country_code":"US","type":"education","lineage":["https://openalex.org/I91045830"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Delbracio, Mauricio","raw_affiliation_strings":["Texas A&M University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Texas A&M University","institution_ids":["https://openalex.org/I91045830"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102759044","display_name":"Hossein Talebi","orcid":"https://orcid.org/0000-0002-5962-2563","observed_orcids":["https://orcid.org/0000-0002-5962-2563"]},"institutions":[{"id":"https://openalex.org/I145311948","display_name":"Johns Hopkins University","ror":"https://ror.org/00za53h95","country_code":"US","type":"education","lineage":["https://openalex.org/I145311948"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Talebi, Hossein","raw_affiliation_strings":["Johns Hopkins University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Johns Hopkins University","institution_ids":["https://openalex.org/I145311948"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004716468","display_name":"Vishal M. Patel","orcid":"https://orcid.org/0000-0002-5239-692X","observed_orcids":["https://orcid.org/0000-0002-5239-692X"]},"institutions":[{"id":"https://openalex.org/I145311948","display_name":"Johns Hopkins University","ror":"https://ror.org/00za53h95","country_code":"US","type":"education","lineage":["https://openalex.org/I145311948"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Patel, Vishal M.","raw_affiliation_strings":["Johns Hopkins University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Johns Hopkins University","institution_ids":["https://openalex.org/I145311948"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5002085979","display_name":"Peyman Milanfar","orcid":"https://orcid.org/0000-0003-1455-7662","observed_orcids":["https://orcid.org/0000-0003-1455-7662"]},"institutions":[{"id":"https://openalex.org/I145311948","display_name":"Johns Hopkins University","ror":"https://ror.org/00za53h95","country_code":"US","type":"education","lineage":["https://openalex.org/I145311948"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Milanfar, Peyman","raw_affiliation_strings":["Johns Hopkins University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Johns Hopkins University","institution_ids":["https://openalex.org/I145311948"]}]}],"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":16,"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":0.9994999766349792,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","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/T10036","display_name":"Advanced Neural Network Applications","score":9.999999747378752e-05,"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/T11689","display_name":"Adversarial Robustness in Machine 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"}}],"keywords":[{"id":"https://openalex.org/keywords/latent-diffusion-models","display_name":"latent diffusion models","score":1.0},{"id":"https://openalex.org/keywords/model-scaling","display_name":"model scaling","score":0.7910000085830688},{"id":"https://openalex.org/keywords/sampling-efficiency","display_name":"sampling efficiency","score":0.9879999756813049},{"id":"https://openalex.org/keywords/text-to-image-generation","display_name":"text-to-image generation","score":0.9470000267028809},{"id":"https://openalex.org/keywords/diffusion-samplers","display_name":"diffusion samplers","score":0.9660000205039978},{"id":"https://openalex.org/keywords/knowledge-distillation","display_name":"knowledge distillation","score":0.9449999928474426},{"id":"https://openalex.org/keywords/generative-models","display_name":"generative models","score":0.4309999942779541}],"concepts":[{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.6963074803352356},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.5811004042625427},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.4639863967895508},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.33231592178344727},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3298899531364441},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2553938627243042},{"id":"https://openalex.org/C97355855","wikidata":"https://www.wikidata.org/wiki/Q11473","display_name":"Thermodynamics","level":1,"score":0.17357569932937622},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.16504329442977905},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.08473002910614014}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2404.01367","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2404.01367","pdf_url":"https://arxiv.org/pdf/2404.01367","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2404.01367","doi":"https://doi.org/10.48550/arxiv.2404.01367","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2404.01367","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2404.01367","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2404.01367","pdf_url":"https://arxiv.org/pdf/2404.01367","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G7555788939","display_name":"CAREER: Seeing Through Atmospheric Turbulence: Image Restoration and Understanding using Deep Convolutional Neural Networks","funder_award_id":"2045489","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4393924573.pdf","grobid_xml":"https://content.openalex.org/works/W4393924573.grobid-xml"},"referenced_works_count":3,"referenced_works":["https://openalex.org/W4379258791","https://openalex.org/W4391709768","https://openalex.org/W4402427382"],"related_works":["https://openalex.org/W4406549391","https://openalex.org/W4415337607","https://openalex.org/W4413144994","https://openalex.org/W4417297081","https://openalex.org/W4416550922","https://openalex.org/W4407365223","https://openalex.org/W4414579597","https://openalex.org/W4406273065","https://openalex.org/W6902324683","https://openalex.org/W7164822207"],"abstract_inverted_index":{"We":[0],"study":[1,104],"the":[2,34,107,110,141],"scaling":[3,145],"properties":[4],"of":[5,31,36,43,55,109,143],"latent":[6],"diffusion":[7,32,58,116],"models":[8,90],"(LDMs)":[9],"with":[10],"an":[11,62],"emphasis":[12],"on":[13],"their":[14,93],"sampling":[15,29,44,70,74],"efficiency.":[16],"While":[17],"improved":[18],"network":[19],"architecture":[20],"and":[21],"inference":[22,87,157],"algorithms":[23],"have":[24],"shown":[25],"to":[26,105,131,151],"effectively":[27],"boost":[28],"efficiency":[30,45,71],"models,":[33,59,124],"role":[35],"model":[37,67],"size":[38,68],"--":[39,46],"a":[40,79,85],"critical":[41],"determinant":[42],"has":[47],"not":[48],"been":[49],"thoroughly":[50],"examined.":[51],"Through":[52],"empirical":[53],"analysis":[54],"established":[56],"text-to-image":[57],"we":[60,101],"conduct":[61],"in-depth":[63],"investigation":[64],"into":[65],"how":[66],"influences":[69],"across":[72],"varying":[73],"steps.":[75],"Our":[76],"findings":[77,112,135],"unveil":[78],"surprising":[80],"trend:":[81],"when":[82],"operating":[83],"under":[84],"given":[86],"budget,":[88],"smaller":[89],"frequently":[91],"outperform":[92],"larger":[94],"equivalents":[95],"in":[96],"generating":[97],"high-quality":[98],"results.":[99],"Moreover,":[100],"extend":[102],"our":[103],"demonstrate":[106],"generalizability":[108],"these":[111],"by":[113],"applying":[114],"various":[115],"samplers,":[117],"exploring":[118],"diverse":[119],"downstream":[120],"tasks,":[121],"evaluating":[122],"post-distilled":[123],"as":[125,127],"well":[126],"comparing":[128],"performance":[129],"relative":[130],"training":[132],"compute.":[133],"These":[134],"open":[136],"up":[137],"new":[138],"pathways":[139],"for":[140],"development":[142],"LDM":[144],"strategies":[146],"which":[147],"can":[148],"be":[149],"employed":[150],"enhance":[152],"generative":[153],"capabilities":[154],"within":[155],"limited":[156],"budgets.":[158]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":8},{"year":2024,"cited_by_count":7}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2024-04-05T00:00:00"}