I-AMD GAIA: Kuyini, ukuthi isebenza kanjani, kanye nendlela yokusebenzisa i-AI yendawo nge-Ryzen

  • I-AMD GAIA iyisixazululo somthombo ovulekile sokusebenzisa i-AI ekhiqizayo endaweni kuma-PC anamandla e-Ryzen AI.
  • Thatha ithuba le-NPU ne-iGPU ukuze wandise ukusebenza futhi uvikele ubumfihlo.
  • Inikeza izindlela ezixubile nezijwayelekile, imidwebo nezixhumanisi zomugqa womyalo, kanye nezakhiwo ezinwebekayo.
Mayka Jimenez

Imizuzu ye-9

Iyini i-AMD GAIA-3

Eminyakeni yamuva, i- Intelligence yokufakelwa Iyekile ukuba umqondo oqhelile futhi isihlanganiswe nezimpilo zansuku zonke zabasebenzisi nezingcweti. Abantu abaningi ngokwengeziwe bafuna izixazululo zokusebenza namamodeli olimi athuthukile, aziwa ngokuthi I-LLM (Amamodeli Olimi Olukhulu), ngendlela yangasese nephumelelayo. Ukuncika kwamafu kanye nokukhathazeka okuyimfihlo kubangele intshisekelo yokusebenzisa lawa ma-AI anamandla ngokuqondile kumakhompyutha womuntu.

Kulokhu kuvela I-AMD GAIA, okukodwa kokubheja kwakamuva futhi okunamandla kwalabo abafuna ukuhlanganisa amandla ezingxenyekazi zekhompuyutha zasendaweni kanye nokusebenza kalula namamodeli e-AI akhiqizayo. Kuthuthukiswe ngu AMD njengesinyathelo se umthombo ovulekile, I-GAIA ithembisa ukuguqula ukusetshenziswa kwe I-Neural Processing Unit (NPU) kumaprosesa Ryzen AI futhi unikeze ngendawo lapho ubumfihlo, ukusebenza kahle kanye nokwenza kube ngokwakho kufinyeleleka khona.

Iyini i-AMD GAIA futhi kungani ibalulekile?

I-GAIA yiyona Isixazululo sesoftware somthombo ovulekile esakhiwe yi-AMD ngokuqondile Faka futhi usebenzise amamodeli we-LLM endaweni, besebenzisa amandla amaphrosesa Ryzen AI. Ngokusabalala kwemisebenzi ye-AI, abasebenzisi abaningi bafuna izindlela zokufinyelela amamodeli asezingeni eliphezulu ngaphandle kokwesaba ukwabelana ngedatha ebucayi emafini futhi ngaphandle kokuthembela ekuxhumekeni unomphela.

Okwenza i-GAIA ihluke akuyona nje imvelo yayo umthombo ovulekile futhi mahhala, kodwa yayo Amandla okusebenzisa i-NPU ehlanganisiwe ochungechungeni lwe-Ryzen AI 300. Lokhu kwenza kube nokwenzeka ukusebenzisa ama-chatbots, abasizi, nama-ejenti ayinkimbinkimbi e-AI ngqo ku-PC yakho, ngobumfihlo obukhulu nokubambezeleka okuphansi kunezinye izindlela eziku-inthanethi. Iphinde igqame ngendlela yayo eyi-hybrid, njengoba ikwazi ukuhlanganisa i-NPU kanye ne I-GPU ehlanganisiwe (iGPU) ngokusebenza okuphezulu nakakhulu, ukuzivumelanisa nezidingo zabo bobabili abasebenzisi basekhaya nonjiniyela abanolwazi.

Ukuze ujule endleleni yokusebenzisa ngokugcwele amakhono we- I-Sysinternals suite ku-Windows futhi uphathe kangcono izinsiza zakho nge-AI, sincoma ukuthi uhlole amanye amathuluzi ahambisana nokuphathwa kwezingxenyekazi zekhompuyutha nesofthiwe kukhompyutha yakho.

Izici ezibalulekile ze-AMD GAIA

Ikhodi kanambambili.

Ukuthuthukiswa kwe-GAIA kusabela esidingweni esikhulayo se sebenzisa amamodeli e-AI endaweni. Ekwakheni nasekusebenzeni kwayo, isoftware ihlanganisa izici ezimbalwa ezithuthuke kakhulu kulo mkhakha:

  • Ukucutshungulwa kwe-LLM yendawo: Ikuvumela ukuthi usebenzise amamodeli olimi ngokusebenza kahle okuphezulu ngqo kumadivayisi e-Windows, ngaphandle kwesidingo sefu.
  • Ukuthuthukisa ihadiwe ye-AMD: Amandla ayo aphezulu afinyelelwa nge Uchungechunge lwe-AMD Ryzen AI 300 sibonga inhlanganisela ye-NPU ne-iGPU, nakuba futhi inemodi evamile kwamanye amakhompyutha.
  • Ubumfihlo obuthuthukisiweUkusebenzisa i-AI endaweni kunciphisa ubungozi bokuvuza noma ukwabelana ngedatha yomuntu siqu ngaphandle kwemvume.
  • Ukubambezeleka okuphansiNgokususa ukuxhumana namaseva angaphandle, kokubili impendulo ye-AI nolwazi lomsebenzisi kuyashesha kakhulu.
  • Izakhiwo ezinwebekayoI-GAIA isekelwe ohlakeni oluvulekile olusiza ukudalwa nokwenza ngendlela oyifisayo ama-ejenti athile kanye namacala okusebenzisa.
  • Isixhumi esibonakalayo esikabili: Inikeza kokubili a isikhombimsebenzisi sokuqhafaza (i-GUI) njengekhonsoli yomugqa womyalo (i-CLI), evumelanisa namaphrofayili ahlukene wabasebenzisi.

Iphuzu elingeziwe yi- Ukusekela ubuchwepheshe bokubuyiswa kolwazi (i-RAG)., enikeza ama-agent e-GAIA ngamakhono athuthukile okusesha ulwazi, isizathu, ukuhlela, nokusebenzisa amathuluzi angaphandle, okudala ukuzizwisa kwengxoxo okucebile kakhulu.

Ubani ongazuza ku-GAIA futhi yiziphi izidingo?

I-GAIA ihloselwe ikakhulukazi abasebenzisi abane- Ikhompyutha ephathekayo noma ideskithophu ene-AMD Ryzen AI 300 series processors. Lawa ma-chips ahlanganisa i-NPU enika amandla ukusheshisa okuthile kokusebenza kobuhlakani bokwenziwa. Uma uhlangabezana nale mfuneko, uzokwazi ukusebenzisa leli thuba imodi eyi-hybrid kusukela ku-GAIA, inketho ehlinzeka ngokusebenza okuphambili kakhulu ngokuhlanganisa i-NPU ne-GPU ehlanganisiwe kuphrosesa efanayo.

Kodwa-ke, i-AMD iphinde yacabanga ngalabo abangenawo la ma-processor. Yingakho kuhlanganisa imodi ejwayelekile, ngokusekelwe ku-backend ye-Ollama, ekuvumela ukuthi usebenzise okunye ukusebenza kwe-GAIA kwamanye amakhompyutha e-Windows, nakuba ukusebenza kungeke kuphakame njengokumiswa kwe-AMD yomdabu.

I-Los ubuncane nezidingo ezinconywayo ukuze i-GAIA isebenze kahle yilokhu okulandelayo:

  • Umseshi: Uchungechunge lwe-AMD Ryzen AI 300 (luyadingeka kumodi ye-hybrid).
  • Imemori ye-RAM: Okungenani u-16 GB, no-32 GB olungele imithwalo enzima kakhulu.
  • Isistimu yokusebenza: Windows 11 Ikhaya noma i-Pro, ibuyekeziwe.
  • Abashayeli: Umshayeli ubuyekezelwe I-AMD Radeon 890M (iGPU) kanye nabashayeli abahambisana ne-NPU.

Amamodi okusebenza: Imodi yeHybrid kanye nemodi ejwayelekile

I-AMD GAIA ihlukaniswa ngokunikeza izindlela ezimbili eziyinhloko zokusebenza, ezivumelanisa nehadiwe yomsebenzisi:

  • Imodi yeHybrid: Uma unekhompyutha efakwe Uchungechunge lwe-AMD Ryzen AI 300, le modi izosebenzisa kokubili i-NPU ne-GPU ehlanganisiwe kanyekanye. Umphumela uba a Isebenzisa amamodeli e-AI ngokushesha, ngempumelelo, nangokusetshenziswa okuphezulu kwehadiwe.
  • Imodi ejwayelekile: Idizayinelwe amakhompyutha angenayo i-Ryzen AI, ithembele kuyo U-Ollama njengemuva lokusebenzisa amamodeli e-LLM ku-Windows 11. Ukusebenza kuphansi, kodwa kukuvumela ukuthi uzame i-GAIA ezinhlobonhlobo zamadivayisi.

Zombili izindlela ziqinisekisa ukuthi abasebenzisi abanezingxenyekazi zekhompuyutha ezisezingeni eliphezulu nalabo abanamathuluzi anesizotha bangahlola, bafunde, futhi bathuthukise izinhlelo zokusebenza ezikhiqizayo ze-AI besekhaya.

Indlela yokufaka i-AMD GAIA igxathu negxathu

Ukufakwa kalula kungenye yamaphuzu aqinile e-GAIA. I-AMD inikeza izifaki ezimbili ezahlukene, kuya ngokuthi ikhompyutha yakho inochungechunge lwe-Ryzen AI 300 noma cha:

  1. I-GAIA Hybrid Installer: Idizayinelwe futhi yenzelwe izinhlelo ze-Ryzen AI-powered, isebenzisa ngokugcwele ukusheshisa kwe-NPU ne-iGPU ehlanganisiwe.
  2. I-GAIA Generic InstallerKuwo wonke amanye ama-Windows PC, sebenzisa i-Ollama backend ukuze uphathe amamodeli amakhulu olimi.

Izinyathelo zokufaka i-GAIA zilula futhi ziyefana kuzo zombili izimo:

  1. Landa isifaki esifanele kusuka endaweni esemthethweni ye-AMD GAIA ku-GitHub, ukukhetha leyo evumelana nokucushwa kwehadiwe yakho.
  2. Vula ifayela elilandiwe bese usebenzisa isifaki .exe esihambisanayo.
  3. Si IWindows ithola isexwayiso sokuvikeleka, khetha okuthi 'Olunye ulwazi' bese uthi 'Gijima noma kunjalo' ukuze uqhubeke.
  4. Inqubo yokufaka ngokuvamile ithatha phakathi Imizuzu engu-5 no-10, kuye ngesivinini soxhumano lwakho lwe-inthanethi.
  5. Lapho usuqedile, uzothola izinqamuleli kuzo zombili i-GUI (GAIA-GUI) kanye ne-CLI (GAIA-CLI) kudeskithophu yakho ukuze uqalise ukusebenzisa isofthiwe.

Ukukhipha i-GAIA uma ufuna ukuyikhipha ohlelweni lwakho nakho kulula: vele uvale noma yiziphi izimo ezivuliwe zohlelo lokusebenza, susa ifolda eyinhloko ye-AppData, susa amafolda ezifanekiso ku-.cache, bese ususa izinqamuleli kudeskithophu yakho.

I-GAIA-GUI ne-GAIA-CLI: izindlela ezimbili zokusebenzisana

I-AMD icabange ngabaqalayo nabasebenzisi abangochwepheshe, ngakho-ke i-GAIA ingasetshenziswa ngezindlela ezimbili ezihlukene:

  • I-Graphical User interface (GUI): Kuyinketho efinyeleleka kakhulu. Chofoza kabili nje isithonjana se-GAIA-GUI kudeskithophu yakho. Uma uyisebenzisa okokuqala, kungase kuthathe isikhathi ukuqalisa, njengoba izodawuniloda amamodeli olimi kungxenyekazi ye-Hugging Face. Uma noma iyiphi imodeli idinga ukhiye we-API, uhlelo lokusebenza luzokwazisa ngalo esikrinini.
  • Ulayini womyalo (CLI): Uma ungumsebenzisi othuthukile, ungafinyelela zonke izici nezilungiselelo ze-GAIA kutheminali ngokusebenzisa umyalo gaia-cli -h ukuze uthintane nezinketho ezitholakalayo.

Lokhu kuvumelana nezimo kwenza i-GAIA ibe yisixazululo esikhangayo kubo bobabili abafuna ukuhlola okungenazinkinga nabathuthukisi abafuna ukulawula okugcwele phezu kokwenziwa nokuhlelwa kwama-ejenti we-AI.

Ama-ejenti, ukunwebeka kanye nohlelo lwe-RAG

Esinye sezici eziphawuleka kakhulu ze-GAIA ukuthi uhlelo lwe-ejenti ehlakaniphile. Lezi Ama-LLM agents Angasebenza ngemizuzu futhi yakhelwe imisebenzi eyahlukene, kusukela ekusekelweni kwengxoxo okuyisisekelo kuye emisebenzini eyinkimbinkimbi kakhulu esebenzisa i-RAG ukubuyisa ulwazi, isizathu, ukuhlela imisebenzi, nokusebenzisa amathuluzi angaphandle.

Okwamanje, i-GAIA ihlanganisa okungenani ama-ejenti amane amakhulu futhi iyaqhubeka nokwandisa ikhathalogi yayo ngenxa yokubambisana komphakathi womthombo ovulekile. Ukwengeza, i-architecture ivumela abasebenzisi noma abathuthukisi ukuthi bakhe futhi bengeze ama-ejenti abo, ngaleyo ndlela benze indawo ezungezile ibe nezidingo zephrojekthi ngayinye.

Uhlelo lwe-Augmented Information Retrieval (AIG) alukhulisi nje kuphela izinga lezimpendulo ze-AI, kodwa futhi luvumela ama-agent ukuthi afinyelele imithombo yedatha yangaphandle futhi enze imisebenzi eyinkimbinkimbi, efana nokuhlanganiswa kolwazi noma ukuhlanganiswa nama-API ezinkampani zangaphandle.

Izinzuzo zokusebenzisa i-LLM endaweni nge-GAIA

Umsebenzisi usebenzisa ikhibhodi ephathekayo.

Amandla okusebenzisa amamodeli e-LLM endaweni nge-GAIA avula umnyango wenombolo izinzuzo ezibonakalayo:

  • Ingqikithi yobumfihlo: Ngokusebenza endaweni, izingxoxo zakho nedatha yomuntu siqu ihlala kudivayisi yakho, ivimbela ukudluliselwa okungafuneki emafini.
  • Ukubambezeleka okuphansi: Ukusebenzisana ne-AI kuyashesha futhi kulula, njengoba kungekho ukubambezeleka okubangelwa ukuxhumana namaseva angaphandle.
  • Ukusebenza okulungiselelwe: Isofthiwe ithuthukiswe ukuze isebenzise ngokugcwele amakhono e-NPU futhi, ngokuzikhethela, i-GPU ehlanganisiwe, ukuzuza ulwazi lwezinga lokuqala.
  • Ukuzivumelanisa nezimo nokwenza ngokwezifiso: Amathuba okudala nokushintsha ama-ejenti noma ukwengeza izici ezintsha makhulu kakhulu kunamasevisi avaliwe noma asekelwe emafini.

I-GAIA ibekwe njengesisetshenziswa esibalulekile sabathuthukisi be-AI, amabhizinisi akhathalela ubumfihlo, kanye nabashisekeli bezobuchwepheshe abafuna ukuyisa i-AI ekhiqizayo ezingeni elilandelayo.

I-GAIA's News and Evolution ngo-2025

Kuwo wonke u-2025, i-GAIA iye yabuyekezwa ngezici ezintsha eziyibeka phezulu engxenyeni yayo. Ezinye zentuthuko ezibaluleke kakhulu zihlanganisa:

  • Ukusekelwa okulungiselelwe kwe-NVIDIA Tensor Cores, ukwandisa ububanzi bezingxenyekazi zekhompuyutha ezihambisanayo nokwenza ukusheshisa kwe-AI kusetshenziswe kumakhompyutha ane-NVIDIA GPU.
  • Ukuthuthukiswa kokuvumelanisa kwamafu, okuvumela ukuhlanganiswa nezinsiza zekhompuyutha njenge-GCP ne-AWS kumaphrojekthi amakhulu adinga izinsiza zangaphandle ezimeni ezithile.
  • Ukuhambisana okunwetshiwe ne-ONNX, okwenza kube lula ukusebenzisana phakathi kwezinhlaka ezihlukene ze-AI nokuvumela ukungenisa/ukuthunyelwa ngaphandle kwamamodeli kalula.

Ngenxa yalezi zibuyekezo, i-GAIA isalokhu iwuphawu phakathi kwabasebenzisi abathuthukile, kuyilapho ivumela futhi ukufinyelela kwentando yeningi nokulula kwalabo abafuna ukuthatha izinyathelo zabo zokuqala ku-AI yendawo.

Kwabafuna ukwazi okwengeziwe noma labo abafuna ukumba bajule, inqolobane esemthethweni ye-GAIA ku-GitHub iqukethe imibhalo enemininingwane, ukufinyelela kubafaki, nezinkundla zokusekela lapho umphakathi wabelana ngamathiphu namasu.

I-GGUF
I-athikili ehlobene:
I-GGUF: Ifomethi eguqula ukusetshenziswa kwamamodeli e-AI endaweni

Engeza njengomthombo okhethwayo ku-Google