{"id":"https://openalex.org/W7148215146","doi":null,"title":"Local Large Language Model Deployment : model size, quantization, and performance trade-offs","display_name":"Local Large Language Model Deployment : model size, quantization, and performance trade-offs","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7148215146"},"language":"en","primary_location":{"id":"pmh:oai:www.theseus.fi:10024/913372","doi":null,"is_oa":true,"landing_page_url":"http://www.theseus.fi/handle/10024/913372","pdf_url":"http://www.theseus.fi/bitstream/10024/913372/2/Soininen_Jere.pdf","source":{"id":"https://openalex.org/S4306400928","display_name":"Theseus (Ammattikorkeakoulujen)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nd","license_id":"https://openalex.org/licenses/cc-by-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"fi=AMK-opinn\u00e4ytety\u00f6|sv=YH-examensarbete|en=Bachelor's thesis|"},"volume":null,"type":"dissertation","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"http://www.theseus.fi/bitstream/10024/913372/2/Soininen_Jere.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5132717728","display_name":"Jere Soininen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Soininen, Jere","raw_affiliation_strings":["University of Applied Sciences , March 2026 ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Applied Sciences , March 2026 ,","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5132717728"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.41280001401901245,"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":0.41280001401901245,"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/T10028","display_name":"Topic Modeling","score":0.29420000314712524,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.07800000160932541,"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/local-large-language-models","display_name":"local large language models","score":0.8330000042915344},{"id":"https://openalex.org/keywords/model-quantization","display_name":"model quantization","score":0.7680000066757202},{"id":"https://openalex.org/keywords/awq","display_name":"AWQ","score":0.9309999942779541},{"id":"https://openalex.org/keywords/qwen3","display_name":"Qwen3","score":0.800000011920929},{"id":"https://openalex.org/keywords/model-size","display_name":"model size","score":0.6349999904632568},{"id":"https://openalex.org/keywords/inference-throughput","display_name":"inference throughput","score":0.6679999828338623},{"id":"https://openalex.org/keywords/energy-efficiency","display_name":"energy efficiency","score":0.8119999766349792},{"id":"https://openalex.org/keywords/inference-latency","display_name":"inference latency","score":0.5460000038146973},{"id":"https://openalex.org/keywords/nvidia-h100","display_name":"NVIDIA H100","score":0.7639999985694885},{"id":"https://openalex.org/keywords/vllm","display_name":"vLLM","score":0.6930000185966492},{"id":"https://openalex.org/keywords/inference-accuracy","display_name":"inference accuracy","score":0.30300000309944153},{"id":"https://openalex.org/keywords/performance-trade-offs","display_name":"performance trade-offs","score":0.5540000200271606},{"id":"https://openalex.org/keywords/on-device-deployment","display_name":"on-device deployment","score":0.5419999957084656},{"id":"https://openalex.org/keywords/data-privacy","display_name":"data privacy","score":0.38199999928474426},{"id":"https://openalex.org/keywords/mmlu-benchmark","display_name":"MMLU benchmark","score":0.48399999737739563},{"id":"https://openalex.org/keywords/gsm8k","display_name":"GSM8K","score":0.7300000190734863},{"id":"https://openalex.org/keywords/hellaswag","display_name":"HellaSwag","score":0.9779999852180481},{"id":"https://openalex.org/keywords/arc-challenge","display_name":"ARC-Challenge","score":0.9810000061988831}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5228000283241272},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.4708000123500824},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.40310001373291016},{"id":"https://openalex.org/C144024400","wikidata":"https://www.wikidata.org/wiki/Q21201","display_name":"Sociology","level":0,"score":0.34619998931884766},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.32739999890327454},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.32659998536109924},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3197000026702881},{"id":"https://openalex.org/C2778137410","wikidata":"https://www.wikidata.org/wiki/Q2732820","display_name":"Government (linguistics)","level":2,"score":0.3140999972820282},{"id":"https://openalex.org/C56739046","wikidata":"https://www.wikidata.org/wiki/Q192060","display_name":"Knowledge management","level":1,"score":0.29750001430511475},{"id":"https://openalex.org/C162118730","wikidata":"https://www.wikidata.org/wiki/Q1128453","display_name":"Actuarial science","level":1,"score":0.2766000032424927},{"id":"https://openalex.org/C39549134","wikidata":"https://www.wikidata.org/wiki/Q133080","display_name":"Public relations","level":1,"score":0.27489998936653137},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.263700008392334},{"id":"https://openalex.org/C19417346","wikidata":"https://www.wikidata.org/wiki/Q7922","display_name":"Pedagogy","level":1,"score":0.2635999917984009},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.259799987077713},{"id":"https://openalex.org/C182306322","wikidata":"https://www.wikidata.org/wiki/Q1779371","display_name":"Order (exchange)","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:www.theseus.fi:10024/913372","doi":null,"is_oa":true,"landing_page_url":"http://www.theseus.fi/handle/10024/913372","pdf_url":"http://www.theseus.fi/bitstream/10024/913372/2/Soininen_Jere.pdf","source":{"id":"https://openalex.org/S4306400928","display_name":"Theseus (Ammattikorkeakoulujen)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nd","license_id":"https://openalex.org/licenses/cc-by-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"fi=AMK-opinn\u00e4ytety\u00f6|sv=YH-examensarbete|en=Bachelor's thesis|"}],"best_oa_location":{"id":"pmh:oai:www.theseus.fi:10024/913372","doi":null,"is_oa":true,"landing_page_url":"http://www.theseus.fi/handle/10024/913372","pdf_url":"http://www.theseus.fi/bitstream/10024/913372/2/Soininen_Jere.pdf","source":{"id":"https://openalex.org/S4306400928","display_name":"Theseus (Ammattikorkeakoulujen)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nd","license_id":"https://openalex.org/licenses/cc-by-nd","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"fi=AMK-opinn\u00e4ytety\u00f6|sv=YH-examensarbete|en=Bachelor's thesis|"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7148215146.pdf","grobid_xml":"https://content.openalex.org/works/W7148215146.grobid-xml"},"referenced_works_count":22,"referenced_works":["https://openalex.org/W2626778328","https://openalex.org/W2747680751","https://openalex.org/W2896457183","https://openalex.org/W2947629474","https://openalex.org/W3030163527","https://openalex.org/W3195577433","https://openalex.org/W4226271314","https://openalex.org/W4308760184","https://openalex.org/W4327810158","https://openalex.org/W4379260375","https://openalex.org/W4387321091","https://openalex.org/W4391670496","https://openalex.org/W4406778835","https://openalex.org/W4407806728","https://openalex.org/W4411403346","https://openalex.org/W4414692173","https://openalex.org/W4414939355","https://openalex.org/W4414946431","https://openalex.org/W4417079115","https://openalex.org/W6910633587","https://openalex.org/W7133225605","https://openalex.org/W7154086658"],"related_works":["https://openalex.org/W7162724405","https://openalex.org/W7213779097","https://openalex.org/W7125417717","https://openalex.org/W7160098174","https://openalex.org/W4417025892","https://openalex.org/W7128957878","https://openalex.org/W7214317047","https://openalex.org/W4412862878","https://openalex.org/W7204259133","https://openalex.org/W7212130659"],"abstract_inverted_index":{"Teko\u00e4ly":[0],"on":[1,43,193],"kehittynyt":[2],"viime":[3],"vuosina":[4],"nopeasti":[5],"ja":[6,30,53,64,73,86,102,108,113,122,129,143,178,185,217],"vakiinnuttanut":[7],"asemansa":[8],"keskeisen\u00e4":[9],"teknologiana":[10],"useilla":[11],"eri":[12],"aloilla.":[13],"Kehityst\u00e4":[14],"ovat":[15,21,56],"vauhdittaneet":[16],"erityisesti":[17,49],"suuret":[18],"kielimallit,":[19],"jotka":[20,214],"saavuttaneet":[22],"laajan":[23],"yleis\u00f6n":[24],"p\u00e4\u00e4asiassa":[25],"pilvipalveluiden,":[26],"kuten":[27],"OpenAI:n":[28],"ChatGPT:n":[29],"Microsoftin":[31],"Copilotin":[32],"kautta.":[33],"N\u00e4it\u00e4":[34],"palveluita":[35],"k\u00e4ytett\u00e4ess\u00e4":[36],"k\u00e4ytt\u00e4jien":[37],"sy\u00f6tt\u00e4m\u00e4":[38],"tieto":[39],"v\u00e4littyy":[40],"palveluntarjoajalle,":[41],"mik\u00e4":[42],"lis\u00e4nnyt":[44],"kiinnostusta":[45],"paikallisia":[46],"ratkaisuja":[47],"kohtaan":[48],"tilanteissa,":[50],"joissa":[51],"tietoturva":[52],"tiedon":[54],"hallinta":[55],"keskeisi\u00e4.":[57],"T\u00e4ss\u00e4":[58],"opinn\u00e4ytety\u00f6ss\u00e4":[59],"tarkastellaan,":[60],"miten":[61],"mallin":[62,159,201],"koko":[63],"kvantisointi":[65,192],"vaikuttavat":[66],"samanaikaisten":[67],"pyynt\u00f6jen":[68],"k\u00e4sittelyyn":[69],"sek\u00e4":[70,89,145],"suorituskyvyn,":[71],"energiatehokkuuden":[72],"tarkkuuden":[74,153],"v\u00e4lisiin":[75],"kompromisseihin":[76],"paikallisesti":[77],"k\u00e4ytt\u00f6\u00f6notetuissa":[78],"suurissa":[79],"kielimalleissa.":[80],"Tutkimus":[81,207],"toteutettiin":[82],"empiirisesti":[83],"Qwen3-malliperheen":[84],"32B-":[85],"8B-malleilla":[87],"BF16-tarkkuudella":[88],"AWQ":[90],"4-bittisin\u00e4":[91],"kvantisoituina":[92],"versioina.":[93],"Mallit":[94],"ajettiin":[95],"vLLM-palvelinmoottorilla":[96],"kahden":[97],"NVIDIA":[98],"H100":[99],"-n\u00e4yt\u00f6nohjaimen":[100],"j\u00e4rjestelm\u00e4ss\u00e4,":[101],"suorituskyky\u00e4":[103],"mitattiin":[104,115],"samanaikaisuustasoilla":[105],"1,":[106],"8":[107],"32":[109,186],"pyynt\u00f6\u00e4.":[110],"P\u00e4\u00e4ttelytehokkuutta,":[111],"energiankulutusta":[112],"laitteistokuormitusta":[114],"ShareGPT-aineistolla":[116],"vLLM":[117],"bench":[118],"serve":[119],"-ty\u00f6kalua":[120],"k\u00e4ytt\u00e4en,":[121],"mallien":[123],"laatua":[124],"arvioitiin":[125],"MMLU-,":[126],"GSM8K-,":[127],"HellaSwag-":[128],"ARC":[130],"Challenge":[131],"-arviointikokonaisuuksilla":[132],"lm-evaluation-harness-ty\u00f6kalun":[133],"avulla.":[134],"Tulokset":[135],"osoittavat,":[136],"ett\u00e4":[137],"AWQ-kvantisointi":[138],"parantaa":[139],"merkitt\u00e4v\u00e4sti":[140],"32B-mallin":[141],"l\u00e4pimenokyky\u00e4":[142],"energiatehokkuutta":[144],"v\u00e4hent\u00e4\u00e4":[146],"muistikaistanleveyden":[147],"k\u00e4ytt\u00f6\u00e4":[148],"yhden":[149,183],"k\u00e4ytt\u00e4j\u00e4n":[150,184,188],"tilanteessa,":[151],"keskim\u00e4\u00e4r\u00e4isen":[152],"heikentyess\u00e4":[154],"vain":[155],"hieman.":[156],"Sen":[157],"sijaan":[158],"koon":[160],"pienent\u00e4minen":[161],"32B:st\u00e4":[162],"8B:hen":[163],"tuottaa":[164],"viel\u00e4":[165],"suuremmat":[166],"tehokkuusedut,":[167],"mutta":[168],"tarkkuus":[169],"heikkenee":[170],"huomattavammin.":[171],"Kaikki":[172],"mallit":[173],"skaalautuvat":[174],"tehokkaasti":[175],"kuorman":[176],"kasvaessa,":[177],"l\u00e4pimenokyky":[179],"kasvaa":[180],"jopa":[181],"20":[182],"samanaikaisen":[187],"v\u00e4lill\u00e4.":[189],"Tulosten":[190],"perusteella":[191],"kustannustehokas":[194],"optimointikeino":[195],"paikalliseen":[196],"kielimallin":[197],"k\u00e4ytt\u00f6\u00f6nottoon,":[198],"kun":[199],"taas":[200],"kokovalinta":[202],"tulisi":[203],"perustaa":[204],"teht\u00e4v\u00e4kohtaisiin":[205],"vaatimuksiin.":[206],"tarjoaa":[208],"k\u00e4yt\u00e4nn\u00f6n":[209],"tietoa":[210],"organisaatioille":[211],"tai":[212],"yksil\u00f6ille,":[213],"harkitsevat":[215],"turvallista":[216],"resurssitehokasta":[218],"paikallista":[219],"kielimalliratkaisua.":[220]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-04-02T00:00:00"}