- I-Windows AI Foundry iyinkundla entsha ehlanganisiwe ye-Microsoft yokuthuthukisa nokuphakela amamodeli e-AI kumadivayisi endawo kanye nasefwini.
- I-Foundry Local inika amandla abathuthukisi ukuthi basebenzise, balungiselele, futhi baphathe amamodeli we-AI azivumelanisa ngokuzenzakalelayo nehadiwe etholakalayo, kuzo zombili iWindows ne-macOS.
- Ukuhlanganiswa ne-Windows ML nokusebenzisana nozakwethu abafana ne-Intel, AMD, Nvidia, ne-Qualcomm kuqinisekisa ukusebenza kahle nokuhambisana kuwo wonke amadivayisi.
- Ama-API amasha namathuluzi okuthuthukisa enza kube lula ukuhlanganiswa kwe-AI ezinhlelweni zokusebenza, okuvumela yonke into kusukela ekwenzeni imodeli ngokwezifiso kuya ezicini ezithuthukisiwe ezifana nosesho lwe-semantic.
I-Microsoft iphinde yaqinisekisa ukuzibophezela kwayo ku-artificial intelligence ngokuthi ukwethula I-Windows AI Foundry kanye nensizakalo Indawo Yokusungula Izinto Zokwakha ngesikhathi somcimbi wayo we-Build 2025, ngaleyo ndlela ihlanganisa isu layo lokubeka i-AI yendawo maphakathi nesipiliyoni sokuthuthukisa ku-Windows nangale kwalokho. Lezi zinhlelo ezintsha ziza ngenhloso yokuhlinzeka abathuthukisi a Indawo evumelana nezimo kakhulu, esebenza kahle futhi eguquguqukayo ukuze udale, uhlole futhi ukhiphe amamodeli e-AI endaweni., eqeda imingcele eminingi evamile.
Ngokufika kweWindows AI Foundry, inkampani ihlose uku lula ukuhlanganiswa kwamamodeli we-AI ezinhlelweni zokusebenza ukwenza okuningi kwezingxenyekazi zekhompuyutha ezitholakala kudivayisi ngayinye nokukhuthaza i-ecosystem evulekile neqinile. Le nhlangano imele ukuzibophezela okucacile ekukhuthazeni ukuqamba okusha ezingeni labasebenzisi nenkampani, Ukunikeza amathuluzi nama-API enza lula kokubili amamodeli wokwakha nokwenza ngokwezifiso umshini wokufunda ukuze asetshenziswe ungaxhunyiwe ku-inthanethi noma emafini.
Ukuzibophezela kweMicrosoft ku-AI yendawo: Iyini i-Windows AI Foundry?
I-Windows AI Foundry Kuwukuvela kwemvelo kwe-Copilot Runtime yangaphambilini, kanye Yethulwa njengenkundla emaphakathi yokuphatha wonke umjikelezo wokuphila wokuthuthukiswa kwe-AI. Lokhu kusuka ekukhethweni kwemodeli nokwenza kahle kuye ekucushweni kahle, ukuqagela, kanye nokuhlanganiswa kwezinhlelo zokusebenza zokugcina, kokubili emagcekeni nangefu le-Azure.
Onjiniyela bangafinyelela amamodeli e-AI yomthombo ovulekile namamodeli obunikazi, kufaka phakathi ethandwa kakhulu emakethe njenge-Ollama, i-Nvidia NIM nezinye izinketho ezikuvumela ukuthi usebenzise ngokunenzuzo amakhathalogi amamodeli ahlukene. Elinye lamaphuzu anamandla awo ikhono lokubona ngokuzenzakalelayo nokujwayela ihadiwe yendawo (CPU, GPU, NPU), incoma futhi ilungiselela amamodeli afaneleka kakhulu esimweni ngasinye ngaphandle kwesidingo sokumisa indawo ngayinye.
La ukusebenzisana nabakhiqizi behadiwe abahamba phambili Okufana ne-Intel, i-AMD, i-Nvidia, ne-Qualcomm kube ukhiye wokuqinisekisa ukuthi i-Windows AI Foundry isebenza kahle, ivumela abathuthukisi ukuthi basebenze ngama-chipset namadivayisi ahlukene, kusukela kuma-laptops angasindi kuya ezindaweni zokusebenza ezinamandla noma ihadiwe ekhethekile efana nama-Copilot+ PC. I-Microsoft inesibopho sokugcina okuncike kusesimweni njalo ukuze kuqinisekiswe ukuhambisana okuphezulu nokusebenza njengoba amaphrosesa ashintsha.
I-Foundry Local: I-AI ngaphandle kokuthembela efwini
Indawo Yokusungula Izinto Zokwakha Igqama njengemojula eyenzelwe labo abafuna ukusebenzisa amamodeli e-AI ngokuphelele endaweni. Ngenxa yomklamo wayo we-cross-platform, abathuthukisi bangafaka amamodeli hhayi kuphela ku-Windows kodwa naku-macOS, okwenza kube lula ukwenza izinhlelo zokusebenza eziyingxube noma izindawo zokusebenza ezixubile ngaphandle kokudela ukuhambisana.
I-Foundry Local inikeza amathuluzi athile afana ne-SDK ne-CLI ukuhlola, ukuthuthukisa, nokusebenzisa amamodeli ngokuqondile endaweni yomsebenzisi. Ungahlola uhlu lwamamodeli ahambisana nehardware yekhompuyutha ngayinye futhi uwasebenzise noma uwenze ngendlela oyifisayo ngokwezidingo zakho. Futhi, ukusizakala Isikhathi sokusebenza se-ONNX, injini ye-Microsoft evulekile yokukhomba esheshisa ukwenziwa kwamamodeli we-AI ezakhiweni ezahlukene.
Lesi sixazululo sihlanganisa ukwesekwa kwe-LoRA (I-Low-Rank Adaptation), indlela evumela amamodeli achazwe kusengaphambili ukuthi afakwe kumasethi edatha yangokwezifiso ngokushesha nangempumelelo, okuyenza ilungele izinkampani noma izinhlangano ezifuna ukujwayelanisa i-AI nezimo ezithile ngaphandle kokusungula amamodeli kusukela ekuqaleni.
Ama-API namandla amasha onjiniyela
IMicrosoft igcizelele kakhulu yenza ukutholwa kwe-AI kube lula ngangokunokwenzekaUkuze wenze lokhu, I-Windows AI Foundry iletha uchungechunge lwama-API alungele ukusetshenziswa igxile emisebenzini evamile efana nokuhlakanipha kombhalo, ukubona isithombe, incazelo yento, nokusesha kwe-semantic. Lawa mathuluzi aqeda isidingo sokuthuthukisa amamodeli wangokwezifiso wemisebenzi eyisisekelo, okuvumela abathuthukisi ukuthi bagxile ekusebenzeni okuhlukile kohlelo lwabo lokusebenza.
Enye inzuzo ebalulekile ukusekelwa kwama-API we ukusesha okuthuthukisiwe nezici zokubuyisa ulwazi, eklanyelwe ngokukhethekile ukuhlanganiswa ekugelezeni komsebenzi kokubuyiswa okungeziwe (i-AR), ukusiza umbuzo nokuhlaziywa kokubili kolwazi oluhlelekile nolungahlelekile.
Konke lokhu kuhambisana nokuhlanganiswa kwe I-Windows ML, isikhathi sokusebenza se-Microsoft se-inference, esenza kube lula ukuthunyelwa kokukhiqiza futhi eqeda isidingo sokupakisha izikhathi zokusebenza noma abashayeli ngaphezu kwalokho. Ngezibuyekezo ezizenzakalelayo, ukuhambisana nokuphepha kuhlala kuqinisekisiwe.
Ukuhambisana, ukuphepha, nokuzizwisa okuhlanganisiwe
I-ecosystem ehlongozwa yiMicrosoft nayo incike ku ukwamukelwa kwamazinga avulekile njenge I-Model Context Protocol (MCP), ukuze abenzeli be-AI namamodeli bakwazi ukuxhumana ngaphandle komthungo nangokuvikelekile kuwo wonke amathuluzi nezinkundla ezahlukene. I-Microsoft izibophezele ekusebenzisaneni kwangempela, isiza ukuguquguquka okuqhubekayo komkhakha kanye nokuhlanganiswa nezixazululo zezinkampani zangaphandle.
Mayelana nokuphepha, inkundla ihlanganisa izici ezithuthukisiwe ezifana nezindawo zokusebenza ezivikelekile (ngokusekelwa kweVirtualization Based Security kanye ne-post-quantum cryptography), kanye namathuluzi okubonakala, ukuqapha ukusebenza nokuvikelwa kosongo.
Inkundla ye-Microsoft ene-Windows AI Foundry kanye ne-Foundry Local ayigcini nje ngokwenza impilo ibe lula kubathuthukisi, kodwa futhi yandisa amathuba amabhizinisi kanye nabasebenzisi bamandla abafuna ukwenza kanjalo. Bopha amandla e-AI ngaphandle kokulahlekelwa ukulawula noma ubumfihlo.
IMicrosoft iqhubekisela phambili ikusasa lokuthuthukiswa kwe-AI, ithuthukisa imodeli lapho ubuhlakani bokwenziwa bungasetshenziswa kahle, kalula futhi ngokuphepha ngqo kumadivayisi abasebenzisi, ngaphandle kokuthembela efwini kuphela. Ukwethulwa kwe-Windows AI Foundry kanye ne-Foundry Local kubeka i-Windows phambili ezinsizakalweni zasendaweni ze-AI, enikeza inkundla eguquguqukayo, evulekile ehambisana ngokugcwele nezidingo ze-ecosystem kanye nabalingani bayo bezobuchwepheshe.
Ngingumshisekeli wezobuchwepheshe oguqule izintshisekelo zakhe "ze-geek" zaba umsebenzi. Ngichithe iminyaka engaphezu kwengu-10 yempilo yami ngisebenzisa ubuchwepheshe obusezingeni eliphezulu kanye nokukitaza ngazo zonke izinhlobo zezinhlelo ngenxa yelukuluku lokufuna ukwazi. Manje sengiqeqeshelwe ubuchwepheshe be-computer nemidlalo yama-video. Lokhu kungenxa yokuthi sekuphele iminyaka engaphezu kwengu-5 ngisebenza ngokubhalela amawebhusayithi ahlukahlukene ezobuchwepheshe nemidlalo yevidiyo, ngenza izindatshana ezifuna ukukunikeza imininingwane oyidingayo ngolimi oluqondakala yiwo wonke umuntu.
Uma unemibuzo, ulwazi lwami lusukela kuyo yonke into ehlobene nesistimu yokusebenza ye-Windows kanye ne-Android yomakhalekhukhwini. Futhi ukuzibophezela kwami kuwe, ngihlala ngizimisele ukuchitha amaminithi ambalwa futhi ngikusize uxazulule noma yimiphi imibuzo ongase ube nayo kulo mhlaba we-inthanethi.