- Funda iimfuno ezibalulekileyo zobugcisa ukuze usebenzise iQwen AI kwiPC yakho.
- Landela amanyathelo aneenkcukacha ukufaka uOllama kunye neDocker ngokufanelekileyo.
- Funda indlela yokuseta ujongano lomzobo olulungele umsebenzisi ukulawula i-AI ekuhlaleni.
Ubukrelekrele bokwenziwa bube sisixhobo esiyimfuneko kwinkalo yetekhnoloji yanamhlanje. Qwen AI, ephuhliswe ngu Alibaba Group, ibalasele njengenye yezona ndlela zintsha nezinamandla. Nangona izisombululo ezininzi ze-AI zifuna uqhagamshelo lwe-intanethi olusisigxina, IQwen AI inokuqhutywa ekuhlaleni kwiWindows 11 ikhompyuter, ukwandisa ulawulo lwabasebenzisi kunye nobumfihlo.
Kwaye ukusebenzisa iQwen AI ekuhlaleni ayiphuculi kuphela ukufikelela okuqhubekayo kwesi sixhobo, kodwa kwakhona Iphelisa imfuneko yokubonisa idatha ebuthathaka kwiinkonzo zefu. Le ndlela ilungele abadwelisi, abaphandi, kunye nabasebenzisi bamandla abafuna ukuguquguquka ngakumbi kumsebenzi wabo wemihla ngemihla. Okulandelayo, siya kuchitha Konke okufuneka ukwazi ukufaka kwaye uqalise ukusebenzisa iQwen AI ngokufanelekileyo.
Iimfuno zokufaka i-Qwen AI ekuhlaleni kwikhompyuter yethu
Ngaphambi kokuba uqale, kufuneka qinisekisa ukuba ikhomputha yakho iyazifezekisa iimfuno zokusebenzisa iQwen AI.
- ukuba iimodeli ezincinci: Kwimizekelo efana naleyo ingaphantsi kwe 4 ibhiliyoni enye yeparameters, ikhomputha enobuncinci 16 GB ye RAM kwaye ngaphandle kweGPU enamandla inokwanela.
- ukuba iimodeli eziphambili ngakumbi: Kwimifuziselo yomthamo ophezulu, njengaleyo inama-20 eebhiliyoni okanye ngaphezulu kweeparamitha, ikhompyutha esebenza kakuhle kakhulu enamakhadi emizobo emininzi iya kufuneka. Ukongezwa kwe-RTX kunye nenkumbulo ephezulu kuneyo 64 GB ye RAM.
Los Izixhobo ze-Hardware nazo zixhomekeke kwimodeli ethile ofuna ukuyifaka.. Iimodeli ezikhaphukhaphu ezifana Qwen2.5-0.5b zilungele imisebenzi esisiseko, ngelixa iinguqulelo eziphambili ezinje Qwen2.5-7b Ifuna uqwalaselo olomelele ngakumbi. Kubalulekile ukuvavanya izixhobo zakho ngaphambi kokuba wenze isigqibo sokuba yeyiphi imodeli oza kuyifaka.
Ukufakela izixhobo eziyimfuneko
Ukuqhuba iQwen AI ekuhlaleni, Kuya kufuneka ufake izixhobo ezibini eziphambili: Ollama y Docker. Inkqubo ichazwe ngezantsi:
- Ollama: Esi sixhobo sasimahla kunye nesivulelekileyo somthombo sikuvumela ukuba usebenzise imifuziselo yolwimi ngqo kwikhompyuter yakho. Ngaba yikhuphele kwi iwebhusayithi yakho esemthethweni, khetha inguqulelo ehambelanayo neWindows kwaye ulandele imiyalelo yesifakeli.
- Idokodo: I-Docker ikuvumela ukuba wenze izikhongozeli zokuqhuba usetyenziso ngokufanelekileyo. Khuphela kwi iphepha layo elisemthethweni, yenza i-akhawunti kwaye ulandele amanyathelo ofakelo kwindlela yakho.
Indlela yokufaka imodeli yeQwen AI
Nge-Ollama kunye ne-Docker emiselweyo, Isinyathelo esilandelayo kukufaka imodeli ye-Qwen AI. Ukusuka kwiwebhusayithi esemthethweni ye-Ollama, fikelela kwifayile ye- icandelo leemodeli kwaye ukhethe imodeli ofuna ukuyisebenzisa. Ngokomzekelo, ukuba unquma ukufaka i-Qwen2.5 kunye ne-parameter ye-0.5b, sebenzisa umyalelo ohambelanayo odweliswe kwiwebhusayithi.
Vula iSitena soMyalelo (CMD) kwikhompyuter yakho, uncamathisele umyalelo kwaye ucinezele faka. Inkqubo inokuthatha imizuzu emininzi, kuxhomekeke kubungakanani bemodeli kunye nesantya soqhagamshelwano lwakho. Nje ukuba ugqityiwe, uya kuba nakho ukuqalisa imibuzo ngokuthe ngqo kwiSitena kunye nomyalelo othile wokubaleka.
Ukumisela ujongano lomzobo
Kwabo bakhetha a amava abonakalayo ngakumbi, kunokwenzeka ukufaka ujongano lwewebhu usebenzisa iDocker. Sebenzisa lo myalelo ulandelayo kwi-CMD Ukwenza isikhongozeli kwiDocker:
docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
Esi sikhongozeli iya kuvula ujongano lomsebenzisi kwibhrawuza yakho, iyafikeleleka kwi izibuko 3000. Apha ungalawula iimodeli zakho ze-AI, uqalise iincoko, kwaye ugcine iimbali zencoko.
Iimodeli ezongezelelweyo ezihambelanayo
Ukongeza kwi-Qwen AI, i-Ollama kunye ne-Docker ivumela ukufakwa kwezinye iimodeli ezifana Llama, Umbambisi y DeepSeek. Ngamnye uneempawu zakhe kunye neenzuzo, ngoko unokuhlola ukhetho olutsha ngokuxhomekeke kwiimfuno zakho ezithile.
Ukuqhuba iimodeli ezongezelelweyo kwandisa amandla enkqubo yakho, kukuvumela ukuba wenze uhlalutyo olunzima ngakumbi kunye nemisebenzi ephezulu ngaphandle kokuxhomekeka kwiinkonzo zangaphandle.
Ukuqhuba iQwen AI kwindawo inikezela ngesisombululo esinamandla kwabo bafuna ukuthatha inzuzo epheleleyo yobukrelekrele bokwenziwa ngaphandle kokuxhomekeka kwi-intanethi okanye kwiiseva zangaphandle. Le nkqubo ayiqinisekisi kuphela ubumfihlo obukhulu, kodwa kwakhona ikuvumela ukuba wenze ngokusesikweni ukusetyenziswa kweemodeli ngokweemfuno ezithile zomsebenzisi ngamnye.
Nangona ifuna ulwazi oluthile lobugcisa, Iziphumo zifanelekile kwabo bafuna ukuthatha ithuba lobu buchwepheshe benguqu.
Ndingumntu othanda itekhnoloji ojike umdla wakhe we "geek" waba ngumsebenzi. Ndichithe ngaphezulu kweminyaka eli-10 yobomi bam ndisebenzisa itekhnoloji yokusika kwaye ndikhenkceza ngazo zonke iintlobo zeenkqubo ngenxa yokufuna ukwazi okumsulwa. Ngoku ndiqeqeshelwe ubugcisa bekhompyutha nakwimidlalo yevidiyo. Oku kungenxa yokuba ngaphezu kweminyaka emi-5 ndibhalela iiwebhusayithi ezahlukeneyo kwitekhnoloji kunye nemidlalo yevidiyo, ndisenza amanqaku afuna ukukunika ulwazi oludingayo ngolwimi oluqondakalayo kuye wonke umntu.
Ukuba unayo nayiphi na imibuzo, ulwazi lwam lusuka kuyo yonke into enxulumene nenkqubo yokusebenza yeWindows kunye ne-Android yeefowuni eziphathwayo. Kwaye ukuzinikela kwam kukuwe, ndihlala ndikulungele ukuchitha imizuzu embalwa kwaye ndikuncede usombulule nayiphi na imibuzo onokuba nayo kweli lizwe le-intanethi.