[{"data":1,"prerenderedAt":1200},["ShallowReactive",2],{"blog-post-nl-\u002Fblog\u002Fpaiton-qwen-image-21-radeon-ai-pro-r9700":3,"blog-posts-sidebar-nl":710},{"id":4,"title":5,"body":6,"categories":690,"date":695,"description":696,"extension":697,"heading":698,"image":699,"meta":700,"navigation":701,"originalUrl":702,"path":703,"seo":704,"slug":705,"socialImage":706,"stem":707,"updated":708,"__hash__":709},"blogNl\u002Fblog\u002Fpaiton-qwen-image-21-radeon-ai-pro-r9700.md","Qwen-Image 2.1 op 1 × Radeon AI PRO R9700: 2048×2048-beelden in 103 seconden",{"type":7,"value":8,"toc":683},"minimark",[9,22,33,70,75,82,98,185,199,215,219,228,251,302,309,355,358,363,382,410,413,417,438,468,495,499,519,538,556,679],[10,11,12,13,17,18,21],"p",{},"Als een beeld op uw eigen werkstation wordt gegenereerd, kunt u de prompt aanpassen en meteen opnieuw proberen zonder het verzoek naar een cloudbeeldendienst te sturen. Met de uitgebrachte ",[14,15,16],"strong",{},"Qwen-Image 2.1 v1.0.2","-bundel kan dat bij 2048 × 2048 pixels met ",[14,19,20],{},"1 × Radeon AI PRO R9700 (32 GB)",".",[10,23,24,25,28,29,32],{},"In drie nieuwe containers duurde een opgewarmd verzoek met het standaardprofiel mediaan ",[14,26,27],{},"103,29 seconden",", van het indienen van het beeldverzoek tot de ontvangst van de PNG. Dat is ongeveer ",[14,30,31],{},"37% minder tijd"," dan de afzonderlijke, historische containerherhaling van v1.0.1 met 165,14 seconden. Het eerste verzoek duurde 113,95 seconden. Dit zijn gemeten releaseresultaten voor deze GPU en werklast, geen belofte voor elke prompt of andere kaart.",[10,34,35,36,39,40,44,45,49,50,62],{},"Het checkpoint, de resolutie en de 40 denoisingstappen blijven gelijk. De tekstencoder, het beeldmodel en de VAE blijven op de GPU: ",[14,37,38],{},"geen CPU-offload en geen VAE-tiling",". Het standaardprofiel wint tijd door ",[41,42,43],"em",{},"na"," de eerste zeven denoisingstappen bewust met lagere precisie te rekenen. Kies ",[46,47,48],"code",{},"exact"," als meer precisiemarge belangrijk is. Beide opties zitten in de openbare container.",[51,52,53],"sup",{},[54,55,61],"a",{"href":56,"ariaDescribedBy":57,"dataFootnoteRef":59,"id":60},"#user-content-fn-guide",[58],"footnote-label","","user-content-fnref-guide","1",[51,63,64],{},[54,65,69],{"href":66,"ariaDescribedBy":67,"dataFootnoteRef":59,"id":68},"#user-content-fn-bench",[58],"user-content-fnref-bench","2",[71,72,74],"h2",{"id":73},"een-volledig-beeldverzoek-gemeten","Een volledig beeldverzoek gemeten",[10,76,77],{},[78,79],"img",{"alt":80,"src":81},"Drie afzonderlijke Qwen-Image 2.1-containerresultaten op één R9700: v1.0.1 historische herhaling 165,14 seconden opgewarmd, v1.0.2 exact 133,74 seconden en v1.0.2 standaard 103,29 seconden. Lager is beter.","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-qwen-image-21\u002Fwarm-request-nl.webp",[10,83,84],{},[41,85,86,87,93,94,21],{},"2048 × 2048 pixels, 40 stappen, guidance 1.0, batchgrootte één, ongewijzigd gebalanceerd MXFP4-checkpoint, één R9700 met 32 GB. De balken tonen de tijd van een volledig opgewarmd HTTP-verzoek tot de PNG-respons. Het gaat om afzonderlijke releaseherhalingen, niet om één gematcht experiment. ",[54,88,92],{"href":89,"rel":90},"https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin\u002Fblob\u002F6586aa618610aed596d2720fbb5768c181bf1011\u002Fmodels\u002FQwen-Image-2.1\u002FBENCHMARKS.md#container-validation",[91],"nofollow","Gepubliceerde metingen"," · ",[54,95,97],{"href":96},"\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-qwen-image-21\u002Fresults.json","grafiekgegevens",[99,100,101,124],"table",{},[102,103,104],"thead",{},[105,106,107,111,115,118,121],"tr",{},[108,109,110],"th",{},"Gepubliceerde containerruntime",[108,112,114],{"align":113},"right","Opgewarmd verzoek",[108,116,117],{"align":113},"Eerste verzoek",[108,119,120],{"align":113},"Piek-VRAM",[108,122,123],{},"Metingen",[125,126,127,145,164],"tbody",{},[105,128,129,133,136,139,142],{},[130,131,132],"td",{},"v1.0.1 historische herhaling",[130,134,135],{"align":113},"165,14 s",[130,137,138],{"align":113},"179,32 s",[130,140,141],{"align":113},"28,34 GiB",[130,143,144],{},"3 processen",[105,146,147,152,155,158,161],{},[130,148,149,150],{},"v1.0.2 ",[46,151,48],{},[130,153,154],{"align":113},"133,74 s",[130,156,157],{"align":113},"147,83 s",[130,159,160],{"align":113},"25,33 GiB",[130,162,163],{},"1 container",[105,165,166,169,174,179,182],{},[130,167,168],{},"v1.0.2 standaard",[130,170,171],{"align":113},[14,172,173],{},"103,29 s",[130,175,176],{"align":113},[14,177,178],{},"113,95 s",[130,180,181],{"align":113},"25,55 GiB",[130,183,184],{},"3 containers",[10,186,187,188,191,192,194,195,198],{},"De tijden voor drie processen zijn ",[14,189,190],{},"medianen","; ",[46,193,48],{}," is één validatierun. Het VRAM-cijfer is het ",[14,196,197],{},"hoogste gemeten gebruik van de volledige GPU"," tijdens de verzoeken van die rij, met een nominale meetinterval van 5 ms. De klok start bij het indienen van het verzoek en stopt na ontvangst van de volledige PNG-respons. Tekstcodering, beeldgeneratie, VAE-decodering en PNG-overdracht zijn inbegrepen. Het downloaden van het model en het opstarten van de server niet.",[10,200,201,202,205,206,208,209],{},"De 37% vergelijkt afzonderlijke herhalingen van gepubliceerde containerreleases. De bron bevat ook gecontroleerde, gematchte tests binnen v1.0.2, maar die hebben een andere steekproef en andere waarden. We voegen deze niet samen tot één gematchte benchmark. De eerdere kwalificatiemeting van ",[14,203,204],{},"103,64 seconden"," is iets anders dan de ",[14,207,27],{}," van de gepubliceerde container hierboven.",[51,210,211],{},[54,212,69],{"href":66,"ariaDescribedBy":213,"dataFootnoteRef":59,"id":214},[58],"user-content-fnref-bench-2",[71,216,218],{"id":217},"begin-op-uw-r9700","Begin op uw R9700",[10,220,221,222,227],{},"Liever een visuele werkomgeving? ",[54,223,226],{"href":224,"rel":225},"https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-studio",[91],"Paiton Studio"," brengt ondersteunde lokale tools voor beeld, video en tekst samen in uw browser. Controleer in de actuele toollijst welke modellen beschikbaar zijn. De commando's hieronder starten de specifieke Qwen-Image 2.1-container die in dit artikel is gemeten.",[10,229,230,231,234,235,238,239,244,245],{},"U hebt Linux, Docker, een werkend AMD GPU-stuurprogramma en minstens ",[14,232,233],{},"30 GiB vrij GPU-geheugen vóór het laden"," nodig. Draai één beeldworker tegelijk. Bij de eerste start worden ongeveer ",[14,236,237],{},"9,33 GB"," checkpointbestanden gedownload en gecontroleerd; latere starts hergebruiken de cache. Vanuit een checkout van de ",[54,240,243],{"href":241,"rel":242},"https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin\u002Ftree\u002Fmain\u002Fmodels\u002FQwen-Image-2.1",[91],"openbare modelhandleiding",":",[51,246,247],{},[54,248,61],{"href":56,"ariaDescribedBy":249,"dataFootnoteRef":59,"id":250},[58],"user-content-fnref-guide-2",[252,253,257],"pre",{"className":254,"code":255,"language":256,"meta":59,"style":59},"language-bash shiki shiki-themes github-light github-dark","git clone https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin.git\ncd paiton-vllm-plugin\ngit checkout 6586aa618610aed596d2720fbb5768c181bf1011\n.\u002Fmodels\u002FQwen-Image-2.1\u002Fserve-docker.sh\n","bash",[46,258,259,275,285,296],{"__ignoreMap":59},[260,261,264,268,272],"span",{"class":262,"line":263},"line",1,[260,265,267],{"class":266},"sScJk","git",[260,269,271],{"class":270},"sZZnC"," clone",[260,273,274],{"class":270}," https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin.git\n",[260,276,278,282],{"class":262,"line":277},2,[260,279,281],{"class":280},"sj4cs","cd",[260,283,284],{"class":270}," paiton-vllm-plugin\n",[260,286,288,290,293],{"class":262,"line":287},3,[260,289,267],{"class":266},[260,291,292],{"class":270}," checkout",[260,294,295],{"class":270}," 6586aa618610aed596d2720fbb5768c181bf1011\n",[260,297,299],{"class":262,"line":298},4,[260,300,301],{"class":266},".\u002Fmodels\u002FQwen-Image-2.1\u002Fserve-docker.sh\n",[10,303,304,305,308],{},"Wacht op het bericht ",[46,306,307],{},"READY",". Dien in een tweede terminal een prompt in en sla de PNG op:",[252,310,312],{"className":254,"code":311,"language":256,"meta":59,"style":59},"python3 models\u002FQwen-Image-2.1\u002Frequest.py \\\n  --prompt 'A neon shop sign that reads \"QWEN IMAGE 2.1\", rainy night, reflections on wet pavement' \\\n  --size 2048 --seed 42 --output outputs\u002Fneon.png\n",[46,313,314,325,335],{"__ignoreMap":59},[260,315,316,319,322],{"class":262,"line":263},[260,317,318],{"class":266},"python3",[260,320,321],{"class":270}," models\u002FQwen-Image-2.1\u002Frequest.py",[260,323,324],{"class":280}," \\\n",[260,326,327,330,333],{"class":262,"line":277},[260,328,329],{"class":280},"  --prompt",[260,331,332],{"class":270}," 'A neon shop sign that reads \"QWEN IMAGE 2.1\", rainy night, reflections on wet pavement'",[260,334,324],{"class":280},[260,336,337,340,343,346,349,352],{"class":262,"line":287},[260,338,339],{"class":280},"  --size",[260,341,342],{"class":280}," 2048",[260,344,345],{"class":280}," --seed",[260,347,348],{"class":280}," 42",[260,350,351],{"class":280}," --output",[260,353,354],{"class":270}," outputs\u002Fneon.png\n",[10,356,357],{},"Voer de client uit vanuit dezelfde repositorymap. Geef elk nieuw verzoek een nieuwe uitvoerbestandsnaam; de client overschrijft geen bestaande afbeelding.",[10,359,360,361,244],{},"Wilt u meer precisiemarge, stop dan de standaardworker en start dezelfde release met het profiel ",[46,362,48],{},[252,364,366],{"className":254,"code":365,"language":256,"meta":59,"style":59},".\u002Fmodels\u002FQwen-Image-2.1\u002Fserve-docker.sh serve --precision-profile exact\n",[46,367,368],{"__ignoreMap":59},[260,369,370,373,376,379],{"class":262,"line":263},[260,371,372],{"class":266},".\u002Fmodels\u002FQwen-Image-2.1\u002Fserve-docker.sh",[260,374,375],{"class":270}," serve",[260,377,378],{"class":280}," --precision-profile",[260,380,381],{"class":270}," exact\n",[10,383,384,385,388,389,392,393,396,397,403,404],{},"De enige containercontrole van dat profiel mat ",[14,386,387],{},"133,74 seconden"," per opgewarmd verzoek. De beeld-API accepteert ook JSON via ",[46,390,391],{},"POST http:\u002F\u002F127.0.0.1:8191\u002Fv1\u002Fimages\u002Fgenerations",". De meegeleverde ",[46,394,395],{},"request.py","-client verwerkt de base64-respons en bewaart de PNG. Deze ",[14,398,399,400],{},"beeld-API staat los van de taalmodelprofielen van ",[46,401,402],{},"paiton serve","; gebruik voor dit model de commando's hierboven.",[51,405,406],{},[54,407,61],{"href":56,"ariaDescribedBy":408,"dataFootnoteRef":59,"id":409},[58],"user-content-fnref-guide-3",[10,411,412],{},"De gekwalificeerde bundel ondersteunt ook transparante PNG's op 1024 of 2048 pixels vierkant en de bewerking van één invoerbeeld naar een uitvoerbeeld van 1024 pixels vierkant. Voor die modi gelden eigen kwaliteits- en prestatiegrenzen. De 103,29 seconden slaan op de tekst-naar-beeldwerklast hierboven.",[71,414,416],{"id":415},"waarom-standaard-sneller-is-en-wanneer-exact-nuttig-is","Waarom standaard sneller is en wanneer exact nuttig is",[10,418,419,420,422,423,425,426,432],{},"De modelgewichten veranderen niet tussen deze runtimeprofielen. In het standaardprofiel gebruiken de verwerking van de tekstprefix en de eerste zeven denoisingstappen het exacte pad. De latere stappen rekenen met lagere precisie. Daardoor duurt de generatie korter, terwijl de eerste stappen, die de compositie sterk bepalen, de rekenprecisie van het ",[46,421,48],{},"-profiel voor hetzelfde gekwantiseerde checkpoint behouden. ",[46,424,48],{}," gebruikt die rekenwijze tijdens alle 40 stappen.",[51,427,428],{},[54,429,61],{"href":56,"ariaDescribedBy":430,"dataFootnoteRef":59,"id":431},[58],"user-content-fnref-guide-4",[51,433,434],{},[54,435,69],{"href":66,"ariaDescribedBy":436,"dataFootnoteRef":59,"id":437},[58],"user-content-fnref-bench-3",[10,439,440,441,447,448,451,452,454,455,458,459,21,462],{},"Het standaardprofiel slaagde voor een ",[14,442,443,444,446],{},"afgebakende kwaliteitscontrole tegenover beelden uit het eigen ",[46,445,48],{},"-profiel",". Er is niet vergeleken met de oorspronkelijke niet-gekwantiseerde BF16-gewichten. De test garandeert evenmin identieke pixels of kwaliteit bij elke prompt. De transparante RGBA-test op 2048 pixels haalde de grens maar nipt: ",[14,449,450],{},"35,02 dB PSNR bij een drempel van 35 dB",". Voor transparant werk of andere toepassingen waarbij meer marge gewenst is, kiest u ",[46,453,48],{},", dat in de containercontrole ongeveer ",[14,456,457],{},"134 seconden"," opgewarmd kostte. De handleiding beschrijft ook het tussenprofiel ",[46,460,461],{},"schedule-int8-11",[51,463,464],{},[54,465,69],{"href":66,"ariaDescribedBy":466,"dataFootnoteRef":59,"id":467},[58],"user-content-fnref-bench-4",[10,469,470,471,473,474,480,481,487],{},"In beide profielen blijven alle modelonderdelen op de GPU. De gemeten pieken van 25,55 GiB (standaard) en 25,33 GiB (",[46,472,48],{},") betreffen de volledige GPU. Het zijn geen downloadgroottes en ze garanderen niet dat een kleinere kaart werkt. De launcher kwalificeert de ",[14,475,476,477],{},"R9700 met RDNA4 ",[46,478,479],{},"gfx1201",", niet elke Radeon of elke GPU met 32 GB.",[51,482,483],{},[54,484,61],{"href":56,"ariaDescribedBy":485,"dataFootnoteRef":59,"id":486},[58],"user-content-fnref-guide-5",[51,488,489],{},[54,490,494],{"href":491,"ariaDescribedBy":492,"dataFootnoteRef":59,"id":493},"#user-content-fn-model",[58],"user-content-fnref-model","3",[71,496,498],{"id":497},"bekijk-de-release-en-de-licenties","Bekijk de release en de licenties",[10,500,501,502,506,507,512,513,518],{},"De ",[54,503,505],{"href":241,"rel":504},[91],"installatiehandleiding"," bevat de actuele container, directe generatie, bewerking en offline instructies. Het ",[54,508,511],{"href":509,"rel":510},"https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin\u002Fblob\u002Fmain\u002Fmodels\u002FQwen-Image-2.1\u002FBENCHMARKS.md",[91],"benchmarkrapport"," toont de metingen per run, het geheugen, de kwaliteitscontroles en de grenzen. De ",[54,514,517],{"href":515,"rel":516},"https:\u002F\u002Fhuggingface.co\u002FEliovpAI\u002FQwen_Image-2.1-MXFP4-Paiton-RDNA4",[91],"Hugging Face-modelkaart"," biedt het checkpoint. De v1.0.2-runtime komt uit de openbare container, niet uit de oorspronkelijke loader op die modelkaart.",[10,520,521,522,525,526,532],{},"Voor de modelgewichten geldt de ",[14,523,524],{},"Qwen Research License",": niet-commercieel onderzoek en evaluatie zijn toegestaan, terwijl commercieel gebruik een aparte upstreamlicentie vereist. Voor de Paiton-runtime en adapter gelden afzonderlijke Apache-2.0-voorwaarden. Controleer beide voordat u het model inzet. De container downloadt de modelgewichten afzonderlijk bij het opstarten.",[51,527,528],{},[54,529,61],{"href":56,"ariaDescribedBy":530,"dataFootnoteRef":59,"id":531},[58],"user-content-fnref-guide-6",[51,533,534],{},[54,535,494],{"href":491,"ariaDescribedBy":536,"dataFootnoteRef":59,"id":537},[58],"user-content-fnref-model-2",[10,539,540,541,545,546,550,551,555],{},"Deze release vormt een extra optie voor lokale beeldgeneratie naast ons ",[54,542,544],{"href":543},"\u002Fnl\u002Fblog\u002Fpaiton-flux2-klein-radeon-ai-pro-r9700","FLUX.2 klein-profiel",". De meetgrens, het model, de instellingen en de licentie zijn anders; de hoofdcijfers vormen dus geen rechtstreekse vergelijking. Ontdek ",[54,547,549],{"href":548},"\u002Fnl\u002Fproducts\u002Fpaiton","Paiton"," of ",[54,552,554],{"href":553},"\u002Fnl\u002Fcontact","neem contact met ons op"," om een AMD-beeldwerklast te bespreken.",[557,558,561,566],"section",{"className":559,"dataFootnotes":59},[560],"footnotes",[71,562,565],{"className":563,"id":58},[564],"sr-only","Footnotes",[567,568,569,625,659],"ol",{},[570,571,573,578,579,586,587,586,594,586,601,586,609,586,617],"li",{"id":572},"user-content-fn-guide",[54,574,577],{"href":575,"rel":576},"https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin\u002Fblob\u002F6586aa618610aed596d2720fbb5768c181bf1011\u002Fmodels\u002FQwen-Image-2.1\u002FREADME.md",[91],"Openbare Qwen-Image 2.1-installatiehandleiding en release-informatie",". ",[54,580,585],{"href":581,"ariaLabel":582,"className":583,"dataFootnoteBackref":59},"#user-content-fnref-guide","Back to reference 1",[584],"data-footnote-backref","↩"," ",[54,588,585,592],{"href":589,"ariaLabel":590,"className":591,"dataFootnoteBackref":59},"#user-content-fnref-guide-2","Back to reference 1-2",[584],[51,593,69],{},[54,595,585,599],{"href":596,"ariaLabel":597,"className":598,"dataFootnoteBackref":59},"#user-content-fnref-guide-3","Back to reference 1-3",[584],[51,600,494],{},[54,602,585,606],{"href":603,"ariaLabel":604,"className":605,"dataFootnoteBackref":59},"#user-content-fnref-guide-4","Back to reference 1-4",[584],[51,607,608],{},"4",[54,610,585,614],{"href":611,"ariaLabel":612,"className":613,"dataFootnoteBackref":59},"#user-content-fnref-guide-5","Back to reference 1-5",[584],[51,615,616],{},"5",[54,618,585,622],{"href":619,"ariaLabel":620,"className":621,"dataFootnoteBackref":59},"#user-content-fnref-guide-6","Back to reference 1-6",[584],[51,623,624],{},"6",[570,626,628,578,633,586,638,586,645,586,652],{"id":627},"user-content-fn-bench",[54,629,632],{"href":630,"rel":631},"https:\u002F\u002Fgithub.com\u002FEliovp-BV\u002Fpaiton-vllm-plugin\u002Fblob\u002F6586aa618610aed596d2720fbb5768c181bf1011\u002Fmodels\u002FQwen-Image-2.1\u002FBENCHMARKS.md",[91],"Qwen-Image 2.1 v1.0.2-benchmarkrapport",[54,634,585],{"href":635,"ariaLabel":636,"className":637,"dataFootnoteBackref":59},"#user-content-fnref-bench","Back to reference 2",[584],[54,639,585,643],{"href":640,"ariaLabel":641,"className":642,"dataFootnoteBackref":59},"#user-content-fnref-bench-2","Back to reference 2-2",[584],[51,644,69],{},[54,646,585,650],{"href":647,"ariaLabel":648,"className":649,"dataFootnoteBackref":59},"#user-content-fnref-bench-3","Back to reference 2-3",[584],[51,651,494],{},[54,653,585,657],{"href":654,"ariaLabel":655,"className":656,"dataFootnoteBackref":59},"#user-content-fnref-bench-4","Back to reference 2-4",[584],[51,658,608],{},[570,660,662,578,667,586,672],{"id":661},"user-content-fn-model",[54,663,666],{"href":664,"rel":665},"https:\u002F\u002Fhuggingface.co\u002FEliovpAI\u002FQwen_Image-2.1-MXFP4-Paiton-RDNA4\u002Fblob\u002Fb4f6bfc00ca13391eb782cb08994a6c7f8bc28bc\u002FREADME.md",[91],"Gepubliceerde Hugging Face-modelkaart, nagekeken revisie",[54,668,585],{"href":669,"ariaLabel":670,"className":671,"dataFootnoteBackref":59},"#user-content-fnref-model","Back to reference 3",[584],[54,673,585,677],{"href":674,"ariaLabel":675,"className":676,"dataFootnoteBackref":59},"#user-content-fnref-model-2","Back to reference 3-2",[584],[51,678,69],{},[680,681,682],"style",{},"html pre.shiki code .sScJk, html code.shiki .sScJk{--shiki-default:#6F42C1;--shiki-dark:#B392F0}html pre.shiki code .sZZnC, html code.shiki .sZZnC{--shiki-default:#032F62;--shiki-dark:#9ECBFF}html pre.shiki code .sj4cs, html code.shiki .sj4cs{--shiki-default:#005CC5;--shiki-dark:#79B8FF}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":59,"searchDepth":277,"depth":277,"links":684},[685,686,687,688,689],{"id":73,"depth":277,"text":74},{"id":217,"depth":277,"text":218},{"id":415,"depth":277,"text":416},{"id":497,"depth":277,"text":498},{"id":58,"depth":277,"text":565},[549,691,692,693,694],"AMD Radeon","Lokale AI","Beeldgeneratie","Qwen","2026-09-23T09:00:00Z","Genereer lokaal Qwen-Image 2.1-beelden van 2048×2048 pixels met 1 × Radeon AI PRO R9700. De Paiton v1.0.2-container mat 103,29 seconden per opgewarmd verzoek, inclusief de PNG-respons.","md","2048 × 2048-beelden in 103 seconden. 1 × Radeon AI PRO R9700.","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-qwen-image-21\u002Fhero-nl.webp",{},true,"https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-qwen-image-21-radeon-ai-pro-r9700","\u002Fblog\u002Fpaiton-qwen-image-21-radeon-ai-pro-r9700",{"title":5,"description":696},"paiton-qwen-image-21-radeon-ai-pro-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-qwen-image-21\u002Fsocial-nl.webp","blog\u002Fpaiton-qwen-image-21-radeon-ai-pro-r9700",null,"astJgmeKwHle4s6ukZbgYMwsXEVeynliMjYhhoOXCJU",[711,713,726,736,748,758,771,780,794,826,838,860,878,897,916,934,951,963,979,994,1006,1015,1023,1038,1050,1061,1072,1082,1095,1105,1118,1129,1139,1150,1159,1171,1182,1191],{"path":703,"title":5,"description":696,"date":695,"slug":705,"image":699,"originalUrl":702,"categories":712},[549,691,692,693,694],{"path":714,"title":715,"description":716,"date":717,"slug":718,"image":719,"originalUrl":720,"categories":721},"\u002Fblog\u002Fpaiton-qwen38-mxfp4-dflash2-r9700","Qwen3.8: 400,7 tok\u002Fs op één R9700 | Paiton","Qwen3.8 op één R9700: 400,7 tok\u002Fs met ROCm 10 en vLLM 0.29, plus publieke 200K\u002F220K-chatprofielen. Benchmarks, beperkingen en startopdrachten.","2026-09-16T07:30:00Z","paiton-qwen38-mxfp4-dflash2-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-qwen38-mxfp4\u002Fupdate-2026-09-19\u002Fhero.webp","https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-qwen38-mxfp4-dflash2-r9700",[549,691,692,722,723,724,725],"AI-inferentie","Inferentie-optimalisatie","Grote taalmodellen","vLLM",{"path":727,"title":728,"description":729,"date":730,"slug":731,"image":732,"originalUrl":733,"categories":734},"\u002Fblog\u002Fpaiton-qwen38-neo-gguf-vllm-r9700","Qwen3.8 GGUF in vLLM: sneller antwoord op één Radeon","Draai de originele NEO CODER MAX GGUF met Paiton in vLLM op een R9700. Bekijk de gemeten responstijden, beeldinvoer en lokale installatie.","2026-09-14T07:30:00Z","paiton-qwen38-neo-gguf-vllm-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-neo-gguf\u002F00-hero-neo-gguf-r9700.webp","https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-qwen38-neo-gguf-vllm-r9700",[549,691,692,735,725],"GGUF",{"path":737,"title":738,"description":739,"date":740,"slug":741,"image":742,"originalUrl":743,"categories":744},"\u002Fblog\u002Fpaiton-minimax-h3-radeon-ai-pro-r9700","MiniMax H3 op Radeon: 15 seconden video met stereogeluid","Paiton genereert lokaal 15 seconden MiniMax H3-video met stereogeluid op één Radeon AI PRO R9700 in 5 min 33 s, met 16,7% minder wachttijd dan stock.","2026-09-09T07:30:00Z","paiton-minimax-h3-radeon-ai-pro-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-minimax-h3\u002F00-featured-minimax-h3-r9700.webp","https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-minimax-h3-radeon-ai-pro-r9700",[549,691,692,745,746,747],"Videogeneratie","MiniMax H3","ComfyUI",{"path":749,"title":750,"description":751,"date":752,"slug":753,"image":754,"originalUrl":755,"categories":756},"\u002Fblog\u002Fpaiton-flux2-klein-radeon-ai-pro-r9700","Lokale FLUX.2 klein op Radeon AI PRO R9700: sneller beelden genereren met minder VRAM","Paiton genereert FLUX.2 klein-beelden van 1024 × 1024 in 1,054 seconden op een R9700, met 16,2% minder generatietijd en 33,4% minder piekallocatie in Torch.","2026-09-07T09:00:00","paiton-flux2-klein-radeon-ai-pro-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-flux2-klein\u002Ffox-paiton.webp","https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-flux2-klein-radeon-ai-pro-r9700",[549,691,692,693,757,747],"FLUX",{"path":759,"title":760,"description":761,"date":762,"slug":763,"image":764,"originalUrl":765,"categories":766},"\u002Fblog\u002Fpaiton-ornith15-radeon-ai-pro-r9700","Ornith 1.5 haalt 44,6 tok\u002Fs op één Radeon AI PRO R9700","Paiton draait Ornith 1.5 35B A3B op één Radeon AI PRO R9700 met 44,63 outputtokens per seconde, 27% sneller en met 21,3% lagere gemodelleerde kosten.","2026-09-05T09:00:00","paiton-ornith15-radeon-ai-pro-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-ornith15\u002F00-featured-ornith15-r9700.webp","https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-ornith15-radeon-ai-pro-r9700",[549,767,691,722,768,769,723,724,725,770],"Kunstmatige intelligentie","GPU-prestaties","Inferentielatentie","Kostenefficiëntie",{"path":772,"title":773,"description":774,"date":775,"slug":776,"image":777,"originalUrl":778,"categories":779},"\u002Fblog\u002Fpaiton-qwen38-radeon-ai-pro-r9700","Paiton: 21% meer Qwen3.8-doorvoer op Radeon AI PRO R9700","Paiton draait AMD's Qwen3.8 27B op één Radeon AI PRO R9700 met 39,77 outputtokens per seconde. Dat levert 21% meer throughput en 17,4% lagere gemodelleerde kosten op.","2026-09-04T09:00:00","paiton-qwen38-radeon-ai-pro-r9700","\u002Fasset\u002Fimages\u002Fblog\u002Fpaiton-r9700\u002F00-featured-paiton-r9700.webp","https:\u002F\u002Feliovp.com\u002Fblog\u002Fpaiton-qwen38-radeon-ai-pro-r9700",[549,767,691,722,768,769,723,724,725,770],{"path":781,"title":782,"description":783,"date":784,"slug":785,"image":786,"originalUrl":708,"categories":787},"\u002Fblog\u002Fai-data-center-power-requirements-gpu-per-megawatt","Stroomvereisten voor AI-datacenters: de GPU-per-MW-illusie","Waarom verschillen GPU-aantallen per megawatt? Lees hoe PUE, piekbelasting, opslag, netwerken en koeling de inzetbare AI-capaciteit bepalen.","2026-07-27T23:52:00","ai-data-center-power-requirements-gpu-per-megawatt","\u002Fasset\u002Fimages\u002Fblog\u002Fai-data-center-power-requirements-gpu-per-megawatt\u002Fgpu-per-megawatt-illusion.webp",[788,789,790,791,792,793],"Alle","AI-infrastructuur","Datacenters","ModFlex","HPC","AMD Helios",{"path":795,"title":796,"description":797,"date":798,"slug":799,"image":800,"originalUrl":801,"categories":802},"\u002Fblog\u002Fpaiton-returns-to-its-diffusion-roots-optimizing-wan2-2-t2v-a14b-on-amd-mi355x","Wan2.2-videogeneratie: Paiton op AMD MI355X","Vergelijk Wan2.2-videogeneratie op AMD MI355X met Paiton en NVIDIA B200 via Diffusers. Lees hoe we diffusiemodellen optimaliseren.","2026-06-10T14:04:04","paiton-returns-to-its-diffusion-roots-optimizing-wan2-2-t2v-a14b-on-amd-mi355x","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fpaitonwan2.webp","https:\u002F\u002Feliovp.com\u002Fpaiton-returns-to-its-diffusion-roots-optimizing-wan2-2-t2v-a14b-on-amd-mi355x\u002F",[788,767,549,803,804,805,806,807,808,809,810,811,812,813,814,815,816,817,818,819,549,820,821,822,823,824,825],"14B","AMD","B200","Benchmarks","Blackwell","Compute","Diffusie","Eliovp","Generatieve AI","GPU","Hardware","Inferentie","Instinct","MI355x","NVIDIA","On-premises","Optimalisatie","Soevereine AI","T2V","Tekst-naar-video","Tuning","Video-generatie","Wan2.2",{"path":827,"title":828,"description":829,"date":830,"slug":831,"image":832,"originalUrl":833,"categories":834},"\u002Fblog\u002Ffrom-the-attic-to-the-front-page-eliovp-recognized-as-a-pioneer-in-chip-optimization-data-center-infrastructure","ElioVP in De Tijd: chipoptimalisatie en datacenters","Lees hoe De Tijd ElioVP belicht, van de oorsprong in chipoptimalisatie tot het werk aan modulaire datacenters en koeling voor hoge vermogensdichtheid.","2026-02-10T20:48:12","from-the-attic-to-the-front-page-eliovp-recognized-as-a-pioneer-in-chip-optimization-data-center-infrastructure","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fphysicalnewspaper.webp","https:\u002F\u002Feliovp.com\u002Ffrom-the-attic-to-the-front-page-eliovp-recognized-as-a-pioneer-in-chip-optimization-data-center-infrastructure\u002F",[788,767,835,836,804,837,835,817],"Modulaire DC","Niet gecategoriseerd","De Tijd",{"path":839,"title":840,"description":841,"date":842,"slug":843,"image":844,"originalUrl":845,"categories":846},"\u002Fblog\u002Fprivacy-is-geen-it-probleem-meer-het-is-een-strategische-prioriteit","AI en privacy: een strategische prioriteit in de Benelux","Privacyrisico's van generatieve AI, vertrouwen, dataopslag en governance. Waarom bedrijven in de Benelux veilige AI strategisch moeten benaderen.","2026-01-29T13:51:11","privacy-is-geen-it-probleem-meer-het-is-een-strategische-prioriteit","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fheaderimage.webp","https:\u002F\u002Feliovp.com\u002Fprivacy-is-geen-it-probleem-meer-het-is-een-strategische-prioriteit\u002F",[788,767,847,836,848,849,850,851,852,853,854,855,810,856,811,857,858,859],"Trending","AI Act","Antropomorfisme","AVG","Benelux","ChatGPT","Cyberbeveiliging","Databeheer","Gegevensbeveiliging","GDPR","Microsoft Copilot","Privacy","Shadow AI",{"path":861,"title":862,"description":863,"date":864,"slug":865,"image":866,"originalUrl":867,"categories":868},"\u002Fblog\u002Fitsme-bij-ons-is-het-its-not-me-en-dit-is-waarom","Waarom wij itsme niet gebruiken: privacy en soevereiniteit","Waarom ElioVP itsme niet gebruikt: onze afwegingen rond identiteitsmetadata, cloudafhankelijkheid, privacy en datasoevereiniteit.","2025-11-27T09:32:14","itsme-bij-ons-is-het-its-not-me-en-dit-is-waarom","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Ffrontimage.webp","https:\u002F\u002Feliovp.com\u002Fitsme-bij-ons-is-het-its-not-me-en-dit-is-waarom\u002F",[788,869,870,871,853,872,873,874,856,875,876,877,858],"AWS","Belgian Mobile ID","CLOUD Act","Datasoevereiniteit","Digitale identiteit","eIDAS","itsme","Liberty Global","MyGov.be",{"path":879,"title":880,"description":881,"date":882,"slug":883,"image":884,"originalUrl":885,"categories":886},"\u002Fblog\u002Ffield-report-the-reality-of-building-agentic-ai-in-2025","Praktijkrapport: de realiteit van Agentic AI bouwen in 2025","Praktijklessen over lokale AI-agents in 2025 gaan in op workflowontwerp, observability, modeltraining, hallucinaties en GPU-geheugenlimieten.","2025-11-25T14:03:39","field-report-the-reality-of-building-agentic-ai-in-2025","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Ffieldreport.webp","https:\u002F\u002Feliovp.com\u002Ffield-report-the-reality-of-building-agentic-ai-in-2025\u002F",[788,767,887,847,888,889,890,891,892,893,894,895,820,896],"Oplossingen","Agentic AI","AI-techniek","AI-strategie","Autonome agenten","Bedrijfs-AI","Lokale LLM","Modelverfijning","AI op locatie","VRAM-optimalisatie",{"path":898,"title":899,"description":900,"date":901,"slug":902,"image":903,"originalUrl":904,"categories":905},"\u002Fblog\u002Fthe-synthetic-unicorn-bubble","De synthetische unicornzeepbel","Een analyse van investeringsrisico’s bij AI-neoclouds: circulaire financiering, infrastructuurclaims, contractvoorwaarden en due diligence.","2025-11-24T19:22:28","the-synthetic-unicorn-bubble","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fsyntheticunicorn.webp","https:\u002F\u002Feliovp.com\u002Fthe-synthetic-unicorn-bubble\u002F",[788,767,847,906,907,908,909,910,911,912,913,914,915],"AI Infrastructure","AI Neocloud","Circulaire financiering","GPU Cloud","Beleggingsrisico's","Opstartwaardering","Synthetische bubbel","Technische analyse","Vaporware","Durfkapitaal",{"path":917,"title":918,"description":919,"date":920,"slug":921,"image":922,"originalUrl":923,"categories":924},"\u002Fblog\u002Fbuilding-the-engine-for-the-ai-race-the-4-month-path-to-nvidia-gb300-nvl72-power","NVIDIA GB300 NVL72: modulair datacenter in vier maanden","Ontdek een modulair datacenterontwerp voor NVIDIA GB300 NVL72, met redundante voeding, hybride koeling en een bouwplanning van vier maanden.","2025-11-20T14:10:19","building-the-engine-for-the-ai-race-the-4-month-path-to-nvidia-gb300-nvl72-power","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fsuperpodmodflexfrontimage.webp","https:\u002F\u002Feliovp.com\u002Fbuilding-the-engine-for-the-ai-race-the-4-month-path-to-nvidia-gb300-nvl72-power\u002F",[788,835,836,925,789,926,927,928,929,930,931,932,933],"150 kW-rack","DLC","Hoge dichtheid","Vloeistofkoeling","Modulair datacenter","NVIDIA Blackwell Ultra","NVIDIA GB300","NVL72","Snelle implementatie",{"path":935,"title":936,"description":937,"date":938,"slug":939,"image":940,"originalUrl":941,"categories":942},"\u002Fblog\u002Fwhy-cuda-translation-wont-unlock-amds-real-potential","CUDA-vertaling versus AMD-gerichte optimalisatie","Waarom CUDA-compatibiliteit niet hetzelfde is als AMD-prestaties: over ROCm, HIP, kerneloptimalisatie en hardwaregerichte afstemming.","2025-11-12T14:48:37","why-cuda-translation-wont-unlock-amds-real-potential","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fchatgpt-image-nov-11-2025-09_16_10-pm-1.webp","https:\u002F\u002Feliovp.com\u002Fwhy-cuda-translation-wont-unlock-amds-real-potential\u002F",[788,767,549,836,943,767,944,945,946,947,948,949,549,950],"AMD MI300X","CUDA-vertaling","FP8","GPU-optimalisatie","High-performance computing","HIP","Kerneltuning","ROCm",{"path":952,"title":953,"description":954,"date":955,"slug":956,"image":957,"originalUrl":958,"categories":959},"\u002Fblog\u002Fpaiton-the-simplest-way-to-supercharge-ai-inference","Paiton: snellere AI-inferentie in uw bestaande stack","Lees hoe Paiton aansluit op bestaande inferentiestacks, met AMD MI300X-benchmarks en vergelijkingen van prestaties per dollar.","2025-11-11T10:31:22","paiton-the-simplest-way-to-supercharge-ai-inference","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fpaiton-powaaah.webp","https:\u002F\u002Feliovp.com\u002Fpaiton-the-simplest-way-to-supercharge-ai-inference\u002F",[788,767,549,722,960,943,770,961,723,949,549,962,725],"AMD Instinct","Hoge throughput","SGLang",{"path":964,"title":965,"description":966,"date":967,"slug":968,"image":969,"originalUrl":970,"categories":971},"\u002Fblog\u002Fstop-overpaying-paiton-mi300x-moe-beats-h200-b200-on-1m-tokens","Paiton MoE-benchmarks: MI300X versus H200 en B200","Vergelijk Qwen3-30B-A3B MoE-benchmarks van MI300X met Paiton, H200 en B200: throughput en kosten per miljoen tokens.","2025-09-26T13:36:18","stop-overpaying-paiton-mi300x-moe-beats-h200-b200-on-1m-tokens","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fhulkvshulkpaitonwins.webp","https:\u002F\u002Feliovp.com\u002Fstop-overpaying-paiton-mi300x-moe-beats-h200-b200-on-1m-tokens\u002F",[788,767,549,972,943,973,723,974,975,976,977,549,978],"AI-benchmarks","Kosten per token","Mixture of Experts","MoE","NVIDIA B200","NVIDIA H200","Qwen3",{"path":980,"title":981,"description":982,"date":983,"slug":984,"image":985,"originalUrl":986,"categories":987},"\u002Fblog\u002Fagentic-ai-but-make-it-local-from-inbox-to-insight-to-action-en","Lokale Agentic AI: van inbox naar actie","Lokale AI-agents zetten e-mails, documenten en beelden om in tickets, rapporten en acties, met modellen op maat van uw gegevens en systemen.","2025-09-16T13:09:00","agentic-ai-but-make-it-local-from-inbox-to-insight-to-action-en","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Ffrontfotoblog.webp","https:\u002F\u002Feliovp.com\u002Fagentic-ai-but-make-it-local-from-inbox-to-insight-to-action-en\u002F",[788,767,887,836,888,988,989,990,991,893,895,820,992,993],"Schadedetectie","Documentverwerking","E-mailautomatisering","Factuurextractie","Ticketautomatisering","Workflowautomatisering",{"path":995,"title":996,"description":997,"date":998,"slug":999,"image":1000,"originalUrl":1001,"categories":1002},"\u002Fblog\u002Fmi300x-fp8-data-parallel-benchmarks-8-64-gpus-h200-left-behind-b200-within-reach","MI300X FP8-benchmarks: GPU-partitionering met Paiton","Bekijk hoe Paiton presteert met Llama 3.1 8B FP8 op gepartitioneerde MI300X-GPU's, vergeleken met NVIDIA H200 en B200.","2025-07-31T13:32:57","mi300x-fp8-data-parallel-benchmarks-8-64-gpus-h200-left-behind-b200-within-reach","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fimage-2-1.webp","https:\u002F\u002Feliovp.com\u002Fmi300x-fp8-data%e2%80%91parallel-benchmarks-8-64-gpus-h200-left-behind-b200-within-reach\u002F",[788,767,549,836,1003,804,805,1004,1005,816,817,549,725],"AI","H200","MI300X",{"path":1007,"title":1008,"description":1009,"date":1010,"slug":1011,"image":1012,"originalUrl":1013,"categories":1014},"\u002Fblog\u002Fapplicable-ai-for-businesses","Toepasbare AI voor bedrijven","Ontdek hoe ElioVP lokale AI voor bedrijfsprocessen bouwt, met modeltraining op maat en automatische schadedetectie voor de logistiek.","2025-07-09T21:35:30","applicable-ai-for-businesses","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fscherm_afbeelding-2025-07-09-om-23.30.17.webp","https:\u002F\u002Feliovp.com\u002Fapplicable-ai-for-businesses\u002F",[788,767,887,888,988,989,990,991,893,895,820,992,993],{"path":1016,"title":1017,"description":1018,"date":1019,"slug":1020,"image":59,"originalUrl":1021,"categories":1022},"\u002Fblog\u002Fintroducing-paitons-free-evaluation-models","Maak kennis met de gratis evaluatiemodellen van Paiton","Test Paiton met gratis evaluatiemodellen voor AMD-GPU's. Vergelijk de prestaties voor tekst, beeldanalyse en beeldgeneratie met uw eigen workloads.","2025-07-07T11:26:13","introducing-paitons-free-evaluation-models","https:\u002F\u002Feliovp.com\u002Fintroducing-paitons-free-evaluation-models\u002F",[788,767,549],{"path":1024,"title":1025,"description":1026,"date":1027,"slug":1028,"image":1029,"originalUrl":1030,"categories":1031},"\u002Fblog\u002Fpaiton-dramatically-faster-startup-and-performance-for-llama-3-1-405b","Llama 3.1 405B: sneller starten met Paiton op MI300X","Bekijk Paiton-benchmarks voor Llama 3.1 405B op acht AMD MI300X-GPU's, met opstarttijd, tensorparallelisme, throughput en latency.","2025-06-12T20:15:23","paiton-dramatically-faster-startup-and-performance-for-llama-3-1-405b","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fservingscreenshot.webp","https:\u002F\u002Feliovp.com\u002Fpaiton-dramatically-faster-startup-and-performance-for-llama-3-1-405b\u002F",[788,767,549,836,722,943,1032,945,1033,1034,1035,549,1036,1037],"Koude start","Grafiekcompilatie","Llama 3.1 405B","LLM-optimalisatie","Opstartlatentie","Tensor-parallellisme",{"path":1039,"title":1040,"description":1041,"date":1042,"slug":1043,"image":1044,"originalUrl":1045,"categories":1046},"\u002Fblog\u002Fpaiton-fp8-beats-nvidias-h200-on-amds-mi300x","Paiton FP8 verslaat NVIDIA's H200 op AMD's MI300X","Vergelijk Paiton op AMD MI300X met NVIDIA H200 voor Llama 3.1 70B FP8: throughput, wachttijd tot het eerste token en latency bij diverse batchgroottes.","2025-06-08T19:12:40","paiton-fp8-beats-nvidias-h200-on-amds-mi300x","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fblognewfp8.webp","https:\u002F\u002Feliovp.com\u002Fpaiton-fp8-beats-nvidias-h200-on-amds-mi300x\u002F",[788,767,549,836,943,1047,892,811,768,769,724,1034,1048,1049],"Koude startoptimalisatie","Model serving","vLLM-optimalisatie",{"path":1051,"title":1052,"description":1053,"date":1054,"slug":1055,"image":1056,"originalUrl":1057,"categories":1058},"\u002Fblog\u002Fmi300x-vs-h200-vs-rx-7900-xtx-vs-tenstorrent-n300s-with-vllm","MI300X, H200, RX 7900 XTX en n300s: vLLM-benchmarks","Vergelijk MI300X, H200, RX 7900 XTX en Tenstorrent n300s met vLLM: throughput, gemodelleerde tokenkosten en hardwarebeperkingen voor Llama 3 8B.","2025-05-09T14:03:58","mi300x-vs-h200-vs-rx-7900-xtx-vs-tenstorrent-n300s-with-vllm","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fcomparisontenstor.webp","https:\u002F\u002Feliovp.com\u002Fmi300x-vs-h200-vs-rx-7900-xtx-vs-tenstorrent-n300s-with-vllm\u002F",[788,767,549,887,836,804,1005,817,1059,1060],"RX7900XTX","tenstorrent",{"path":1062,"title":1063,"description":1064,"date":1065,"slug":1066,"image":1067,"originalUrl":1068,"categories":1069},"\u002Fblog\u002Fclusterpl-empowering-gpu-cluster-investors-with-real-world-financial-insights","ClusterP&L: financiële modellen voor GPU-clusters","Ontdek hoe ClusterP&L kosten, rendement en investeringsscenario's voor GPU-clusters modelleert, met risicoanalyses en exporteerbare rapporten.","2025-05-03T10:52:22","clusterpl-empowering-gpu-cluster-investors-with-real-world-financial-insights","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fcomparisonscenarios.webp","https:\u002F\u002Feliovp.com\u002Fclusterpl-empowering-gpu-cluster-investors-with-real-world-financial-insights\u002F",[788,767,835,887,805,1004,1070,817,1071],"MI325X","P&L-calculator",{"path":1073,"title":1074,"description":1075,"date":1076,"slug":1077,"image":1078,"originalUrl":1079,"categories":1080},"\u002Fblog\u002Fcranking-out-faster-tokens-for-fewer-dollars-amd-mi300x-vs-nvidia-h200","AMD MI300X versus NVIDIA H200: Qwen3-32B met Paiton","Vergelijk Qwen3-32B-benchmarks op AMD MI300X met Paiton en NVIDIA H200, met resultaten voor throughput, latency en hardwarekosten.","2025-05-02T21:10:30","cranking-out-faster-tokens-for-fewer-dollars-amd-mi300x-vs-nvidia-h200","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002F3ac59a73-2466-4422-b7e5-ef2e4a8ca58e.webp","https:\u002F\u002Feliovp.com\u002Fcranking-out-faster-tokens-for-fewer-dollars-amd-mi300x-vs-nvidia-h200\u002F",[788,767,549,1003,804,1004,1081,817,549,725],"MI300",{"path":1083,"title":1084,"description":1085,"date":1086,"slug":1087,"image":1088,"originalUrl":1089,"categories":1090},"\u002Fblog\u002Fpower-meets-precision-high-density-modular-data-center-for-nvidia-nvl-deployments-1-2-mw","Modulaire datacenters voor NVIDIA NVL: 1 tot 2 MW","Ontdek modulaire datacenterontwerpen voor NVIDIA NVL-systemen, met aandacht voor vermogen, vloeistofkoeling, redundantie en uitrolplanning.","2025-05-02T14:09:59","power-meets-precision-high-density-modular-data-center-for-nvidia-nvl-deployments-1-2-mw","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Feliovp_critical-1mw-pod_rev-2_transparent.webp","https:\u002F\u002Feliovp.com\u002Fpower-meets-precision-high-density-modular-data-center-for-nvidia-nvl-deployments-1-2-mw\u002F",[788,835,1091,906,927,792,928,929,1092,1093,932,1094],"1-2MW datacenter","NVIDIA Blackwell","NVIDIA NVL","Precisiekoeling",{"path":1096,"title":1097,"description":1098,"date":1099,"slug":1100,"image":1101,"originalUrl":1102,"categories":1103},"\u002Fblog\u002Fexamining-ai-agents-in-the-medical-field-ai-that-speaks-dicom","AI-agents in de medische wereld: AI die DICOM spreekt","Ontdek een lokale AI-agent die DICOM-gegevens opzoekt en bekijk tests van beeldmodellen met geanonimiseerde medische beelden.","2025-04-11T14:45:48","examining-ai-agents-in-the-medical-field-ai-that-speaks-dicom","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fhealthcareblog-1.webp","https:\u002F\u002Feliovp.com\u002Fexamining-ai-agents-in-the-medical-field-ai-that-speaks-dicom\u002F",[788,767,887,836,1003,804,1104],"Zorg",{"path":1106,"title":1107,"description":1108,"date":1109,"slug":1110,"image":1111,"originalUrl":1112,"categories":1113},"\u002Fblog\u002Feliovp-bv-your-trusted-partner-for-supply-chain-resilience-amidst-new-u-s-tariffs","Amerikaanse invoerheffingen en AI-leveringszekerheid in 2025","Lees ElioVP's visie uit april 2025 op Amerikaanse invoerheffingen en leveringszekerheid voor AI-servers, HPC-systemen en modulaire datacenters.","2025-04-04T10:01:27","eliovp-bv-your-trusted-partner-for-supply-chain-resilience-amidst-new-u-s-tariffs","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Ftariffsshipping.webp","https:\u002F\u002Feliovp.com\u002Feliovp-bv-your-trusted-partner-for-supply-chain-resilience-amidst-new-u-s-tariffs\u002F",[788,847,1003,804,1114,1115,1116,1117],"Invoer","Taiwan","Invoerheffingen","Trump",{"path":1119,"title":1120,"description":1121,"date":1122,"slug":1123,"image":1124,"originalUrl":1125,"categories":1126},"\u002Fblog\u002Fwhy-ai-agents-are-the-future","Waarom AI-agenten de toekomst zijn","Ontdek AI-agents voor ERP, CRM, financiën en klantondersteuning, met praktijkvoorbeelden en een traject van procesanalyse tot pilot en uitrol.","2025-03-23T22:06:59","why-ai-agents-are-the-future","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Ferp2.jpeg","https:\u002F\u002Feliovp.com\u002Fwhy-ai-agents-are-the-future\u002F",[788,767,887,1003,1127,1128],"AI-agenten","ERP",{"path":1130,"title":1131,"description":1132,"date":1133,"slug":1134,"image":1135,"originalUrl":1136,"categories":1137},"\u002Fblog\u002Fthe-rise-of-open-source-ai-model-optimization","De opkomst van open-source AI-modeloptimalisatie","Verken trends in opensource-AI-optimalisatie: kwantisatie, mixture-of-experts-modellen, hardwaregerichte afstemming, RAG en edge-AI.","2025-03-22T20:59:23","the-rise-of-open-source-ai-model-optimization","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Friseofopensource.jpeg","https:\u002F\u002Feliovp.com\u002Fthe-rise-of-open-source-ai-model-optimization\u002F",[788,767,847,1138,804,812,817],"AI-nieuws",{"path":1140,"title":1141,"description":1142,"date":1143,"slug":1144,"image":1145,"originalUrl":1146,"categories":1147},"\u002Fblog\u002Fintroducing-our-benchmarking-tool-powered-by-dstack","Maak kennis met onze benchmarktool, gebouwd op dstack","Ontdek onze benchmarktool met dstack: herhaalbare vLLM-tests, automatische parameterreeksen en prestatierapporten voor lokale GPU's en de cloud.","2025-03-20T14:21:59","introducing-our-benchmarking-tool-powered-by-dstack","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fbenchmarktool.jpeg","https:\u002F\u002Feliovp.com\u002Fintroducing-our-benchmarking-tool-powered-by-dstack\u002F",[788,767,549,1003,804,1148,1149,1005,549],"benchmark","LLM",{"path":1151,"title":1152,"description":1153,"date":1154,"slug":1155,"image":1156,"originalUrl":1157,"categories":1158},"\u002Fblog\u002Foptimizing-qwq-32b-by-qwen-amd-mi300x-vs-nvidia-h200","QwQ-32B optimaliseren (door Qwen): AMD MI300X versus NVIDIA H200","Vergelijk throughput en latency van QwQ-32B op AMD MI300X met Paiton en NVIDIA H200, bij kleine batches en meer gelijktijdige aanvragen.","2025-03-19T21:41:44","optimizing-qwq-32b-by-qwen-amd-mi300x-vs-nvidia-h200","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fpaiton4.jpeg","https:\u002F\u002Feliovp.com\u002Foptimizing-qwq-32b-by-qwen-amd-mi300x-vs-nvidia-h200\u002F",[788,767,549],{"path":1160,"title":1161,"description":1162,"date":1163,"slug":1164,"image":1165,"originalUrl":1166,"categories":1167},"\u002Fblog\u002Feliovp-featured-on-amd-tech-talk-podcast","Eliovp te gast in de AMD Tech Talk-podcast","Beluister Elio Van Puyvelde en Jim Greene in de AMD Tech Talk-podcast over het ontstaan van ElioVP en de hardware- en softwarediensten voor AI.","2025-03-19T07:53:39","eliovp-featured-on-amd-tech-talk-podcast","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Ftechtalkjimgreene.jpeg","https:\u002F\u002Feliovp.com\u002Feliovp-featured-on-amd-tech-talk-podcast\u002F",[788,804,1168,1169,1170],"Jim Greene","Podcast","Tech Talk",{"path":1172,"title":1173,"description":1174,"date":1175,"slug":1176,"image":1177,"originalUrl":1178,"categories":1179},"\u002Fblog\u002Ffurther-optimizing-amd-powered-inference-with-paiton","AMD-inferentie verder optimaliseren met Paiton","Bekijk Paiton-benchmarks voor DeepSeek R1 Distill Llama 8B op AMD MI300X, gericht op throughput en latency bij kleinere batchgroottes.","2025-03-13T06:18:30","further-optimizing-amd-powered-inference-with-paiton","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fpaitonpost3.webp","https:\u002F\u002Feliovp.com\u002Ffurther-optimizing-amd-powered-inference-with-paiton\u002F",[788,767,549,804,1180,1181,1004,1005,1070,549,725],"DeepSeek","H100",{"path":1183,"title":1184,"description":1185,"date":1186,"slug":1187,"image":1188,"originalUrl":1189,"categories":1190},"\u002Fblog\u002Fa-first-look-at-paiton-in-action-deepseek-r1-distill-llama-3-1-8b","Paiton-benchmarks: DeepSeek R1 Distill Llama 3.1 8B","Vergelijk standaard- en Paiton-versies van DeepSeek R1 Distill Llama 3.1 8B op AMD MI300X, met benchmarks voor throughput en latency.","2025-01-31T09:11:02","a-first-look-at-paiton-in-action-deepseek-r1-distill-llama-3-1-8b","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fpaitonpost2.webp","https:\u002F\u002Feliovp.com\u002Fa-first-look-at-paiton-in-action-deepseek-r1-distill-llama-3-1-8b\u002F",[788,767,549,804,1180,1181,1004,1005,1070,549,725],{"path":1192,"title":1193,"description":1194,"date":1195,"slug":1196,"image":1197,"originalUrl":1198,"categories":1199},"\u002Fblog\u002Fai-model-optimization-with-paiton","AI-modeloptimalisatie met Paiton","Lees hoe Paiton modelcompilatie, aangepaste kernels en kernelfusie inzet om AI-inferentie op AMD GPU's te optimaliseren.","2025-01-30T19:53:25","ai-model-optimization-with-paiton","\u002Fasset\u002Fimages\u002Fblog\u002Fimported\u002Fpaitonpost1.webp","https:\u002F\u002Feliovp.com\u002Fai-model-optimization-with-paiton\u002F",[788,767,549,804,1181,1004,1005,1070,549,725],1790198254310]