Kopisha izinhlelo zokusebenza ezinobuhlakani bokwenziwa usebenzisa amaphrojekthi omthombo ovulekile

Ukuhlanganiswa kwe izicelo ngokwezifiso kanye nobuhlakani bokwenziwa Kuvule umnyango wento eyayibonakala njengendaba yesayensi eminyakeni embalwa edlule: ikhono lokulingisa, ukuzivumelanisa, noma ukuphinda usebenze kwezinhlelo zokusebenza ezinamandla usebenzisa amaphrojekthi e-AI omthombo ovulekile. Namuhla, awudingi ukuba yinkampani enkulu yamazwe ngamazwe noma ube ne-supercomputer ukuze uzame lobu buchwepheshe futhi ubuguqule bube yizixazululo zangempela zebhizinisi lakho.

Kulesi sihloko uzothola umhlahlandlela ophelele wokuthi kopisha izinhlelo zokusebenza ezisebenzisa i-AI ngokunesibophoUkuqonda "ukukopisha" njengamakhono okuphinda, imisebenzi, kanye nemiphumela, hhayi njengokuphula ilungelo lobunikazi. Sizobheka ukukhethwa kwamaphrojekthi omthombo ovulekile ongawafaka kukhompyutha yakho noma kudivayisi yakho yeselula, ukuthi angena kanjani ezindaweni zenkampani, nokuthi ithimba elikhethekile lisebenza kanjani Isofthiwe eyenziwe ngokwezifiso, ukuphepha kwe-cyber, izinsizakalo zefu ze-AWS kanye ne-Azure, kanye ne-AI yamabhizinisi kungakusiza ukuthi uzihlanganise ngobungcweti.

Kusho ukuthini ngempela ukuthi “ukukopisha izinhlelo zokusebenza” nge-AI?

Uma sikhuluma ngakho kopisha izinhlelo zokusebenza ngobuhlakani bokwenziwa Asikhulumi ngokulanda ikhodi kusuka kuhlelo lokusebenza lwezentengiselwano bese siyikopisha ngqo, okungaba okungemthetho futhi okungekho emthethweni. Umqondo uwukusebenzisa amaphrojekthi omthombo ovulekile ukuze ukuziphatha okuphindaphindayo kanye nezinsiza esevele ikhona emakethe: abasizi bezwi, abakhiqizi bezithombe, izinjini zokusesha zangaphakathi, ababhali bezithombe abazenzakalelayo, njll.

La maphrojekthi avulekile avumela noma ubani ukuthi hlanganisa indawo yokugcina, uyifake bese uyivumelanisa nezimo ngokwezidingo zabo. Kusukela lapho, ithimba lokuthuthukisa lingakha ungqimba lwe ukusetshenziswa kwesiko okuhambisana nezinqubo zenkampani yakho, izidingo zakho zokuphepha kanye nezinhlelo zakho ezikhona (i-ERP, i-CRM, i-intranet, izizindalwazi, izinsizakalo zefu ze-AWS kanye ne-Azure, njll.).

Ngenxa ye-ecosystem yomthombo ovulekile, manje sekulula ukwakha ama-prototypes e- Ama-ejenti e-AI, abasizi bangaphakathi, noma amathuluzi obuhlakani bebhizinisi ngaphandle kokukhokha imali yelayisensi yokuqala ebiza kakhulu. Kodwa-ke, udinga ukwazi ukuthi ungayikhetha kanjani iphrojekthi efanele, usayizi wengqalasizinda, futhi unakekele ukuphepha kwe-inthanethi kanye nokuthobela imithetho kusukela osukwini lokuqala.

Ezigabeni ezilandelayo sizobona Amaphrojekthi angu-10 abalulekile Lawa mathuluzi akuvumela ukuthi "ukopishe" amakhono avela kuzinhlelo zokusebenza ze-AI ezidumile: ukubhalwa okuzenzakalelayo, ama-chatbot, ukukhiqizwa kwezithombe, izinjini zokusesha zebhizinisi, ukwenziwa kwezwi, abasizi bokuthuthukiswa, nokunye okuningi. Wonke lawa maphrojekthi angatholakala ku-GitHub, afakwe emishinini yakho, futhi, okubaluleke kakhulu, anwetshwe ngokusebenzisa [ukuthuthukiswa/amaphrojekthi/njll.]. isofthiwe yangokwezifiso.

I-Whisper yi-OpenAI: Bhala umsindo njengezinhlelo zokusebenza ezinhle kakhulu

Izinhlelo zokusebenza eziningi ezidumile zithembele ekuguqulweni kwenkulumo ibe umbhalo okunembe kakhulu. I-Whisper yi-OpenAI Ungafinyelela into efanayo kakhulu engqalasizinda yakho. Kuyimodeli yomthombo ovulekile we bhala umsindo ube umbhalo Kutholakala ngezilimi eziningi, kulungele ama-podcast, imihlangano, amakilasi, noma izingcingo zamakhasimende.

Ukufakwa kwe-Whisper okuvamile kuhilela ukusebenzisa i-Python kanye nokuqalisa i- ukufaka i-pip openai-whisper, ngemuva kokwenza i-repository esemthethweni ibe yi-coloning ku- github.com/openai/whisperUkusuka lapho, ungayihlanganisa neziqophi zezwi, amapulatifomu emihlangano yevidiyo, noma izinhlelo zokusebenza zangaphakathi ezisetshenziswa kakade yithimba lakho.

Esimweni sebhizinisi, i-Whisper ifaneleka kahle kakhulu dala amaminithi omhlangano ozenzakalelayobhala phansi ukuqeqeshwa kwangaphakathi noma udale i-corpus yokusesha okungabe sekusetshenziswa izinsizakalo zobuhlakani bebhizinisi namathuluzi afana nalawa I-Power BICabanga nje ukuthi yonke imihlangano yabaphathi ihlelwe futhi isesheka kusetshenziswa ulimi lwemvelo.

Isihluthulelo "sokukopisha" izinhlelo zokusebenza zokubhala ngokuhweba ukusonga i-Whisper nge- ukusetshenziswa kwesikoIsibonisi esilula, ukuphathwa komsebenzisi, isitoreji esiphephile samafayela omsindo nombhalo, ukuhlanganiswa nezinsizakalo zefu ze-AWS ne-Azure, kanye nezinqubomgomo ukuphepha kwe-cyber okulungiselelwe ngokwemboni yakho (ezomthetho, ezempilo, ezezimali, njll.).

Ama-AutoGPT kanye nama-ejenti e-AI azimele ukuze enze imisebenzi ngokuzenzakalela

Uma umgomo wakho uwukulingisa izinhlelo zokusebenza ezihlela futhi zenze imisebenzi cishe ngokuzenzakalelayo, I-AutoGPT Kungenye yamaphrojekthi amakhulu. Kuyi- i-autonomo okuncike kumamodeli e-GPT ukuhlukanisa umgomo ube imisebenzi emincane bese uwenza ngaphandle kokuthi kudingeke ukuthi ulandelele isinyathelo ngasinye.

Le phrojekthi itholakala e- github.com/Significant-Gravitas/Auto-GPT futhi idinga i-Python kanye ne- Ukhiye we-API ngemodeli yolimi ofuna ukuyisebenzisa. Uma isilungisiwe, ungayitshela ukuthi icwaninge ngesihloko, ikhiqize umbiko, ilungiselele isu lokumaketha kokuqukethwe, noma yenze ingxenye yokusebenzisana namathuluzi aku-inthanethi ibe ngokuzenzakalela.

Izinhlelo zokusebenza eziningi zamanje zokukhiqiza kanye nokwenza ngokuzenzakalela ezisekelwe ku-AI zilandela lesi simiso esifanayo: a I-ejenti ye-AI ethola umgomo bese yenza izinqumo zokuwuqeda. Nge-AutoGPT, izinkampani zingakwazi abasizi bangaphakathi besibonelo ukuthi baphatha imisebenzi efana izihloko ezisalungiswa, dala imikhankaso yokumaketha nge-imeyili noma uqoqe ulwazi mayelana nabancintisana nabo.

Kodwa-ke, ukuze kusuke ekuhlolweni kuye emkhiqizweni wangempela, le agent idinga ukugoqwa ngesendlalelo se- isofthiwe yangokwezifiso ebeka imingcele ecacile, iqopha izenzo, futhi ihambisana nezidingo ze ukuphepha kwe-cyberIthimba elihle lokuthuthukisa lingahlanganisa i-AutoGPT namadeshibhodi angaphakathi, imisebenzi yebhizinisi, kanye nezinsizakalo zamafu ze-AWS ne-Azure ukuze zikhule uma kudingeka.

I-GPT4All kanye ne-llama.cpp: ama-chatbot endawo ngaphandle kokuthembela efwini

Esinye isigaba sezinhlelo zokusebenza abantu abaningi abafuna "ukuzikopisha" yilesi Ama-chatbot ohlobo lwe-ChatGPTkodwa kwenziwa endaweni yangakini ngenxa yobumfihlo noma ngezizathu zezindleko. Amaphrojekthi amabili avelele lapha: GPT4All y shayela.cppokukuvumela ukuthi usebenzise amamodeli wombhalo akhiqizayo kukhompyutha noma kuseva yakho.

GPT4All, itholakala ngaphakathi github.com/nomic-ai/gpt4allInikeza isikhombikubona esisebenziseka kalula sokusebenzisa amamodeli e-GPT endaweni. Ilungele ukufakwa abasizi bangaphakathi ezingancikile ezinsizakalweni zangaphandle, into ebaluleke kakhulu ezinkampanini ezisingatha idatha eyimfihlo noma ezifuna ukunciphisa ukuthembela kwazo efwini.

Ngakolunye uhlangothi, shayela.cpp (e github.com/ggerganov/llama.cppIlungiselelwe ukusebenzisa amamodeli omndeni we-LLaMA kahle kakhulu ku-CPU, ngisho nasemishinini emincane. Ikuvumela ukuthi wakhe izinhlelo zokusebenza zengxoxo, abasizi bokusekela bangaphakathi, kanye nama-ejenti e-AI angaxhunyiwe ku-inthanethi, iwusizo kakhulu emibonisweni yamakhasimende noma ezindaweni ezinokuxhumana okulinganiselwe.

Ngalezi zinhlelo ezimbili ungakha izinhlelo zokusebenza eziphinda ulwazi lwengxoxo lwamapulatifomu amakhulu ezentengiselwano, ngenkathi ugcina ukulawula okugcwele ingqalasizinda, ubumfihlo, kanye nokuphepha kwe-inthanethiNgokombono webhizinisi, zingahlanganiswa nezizindalwazi zangaphakathi, izinhlelo zokuthengisa amathikithi, noma ama-intranet ezinkampani ngokusebenzisa isofthiwe yangokwezifiso.

I-PrivateGPT kanye ne-Haystack: izinjini zokusesha kanye ne-QA yamadokhumenti akho

Izinkampani eziningi zifuna "ukukopisha" imisebenzi evela kumathuluzi afana nabasizi bemibhalo noma izinjini zokusesha ezihlakaniphile eziphendula imibuzo mayelana nezincwadi, izinkontileka, noma izinqubomgomo zangaphakathi. Kulesi sigaba, ziyagqama. I-PrivateGPT y i-deepset Haystack.

I-PrivateGPT, itholakala ngaphakathi github.com/imartinez/privateGPT, kuyakuvumela bhala imibuzo bese uthola izifinyezo ezivela kumadokhumenti endawo (Ama-PDF, izinkontileka, imibiko…) ngaphandle kokulayisha noma yini efwini. Ilungele iminyango yezomthetho, izinsizakusebenza zabantu, noma yiliphi iqembu eliphatha ulwazi olubucayi kakhulu ngezimfuneko eziqinile. ukuphepha kwe-cyber.

Umqondo ulula kakhulu: ukopisha indawo yokugcina, ungeze amafayela akho, bese uhlelo lukhiqiza inkomba ongayibuza usebenzisa ulimi lwemvelo. Lokhu kukuvumela ukuthi "ukopishe" okuhlangenwe nakho kohlobo lomsizi wedokhumenti yebhizinisi ngaphandle kokuthembela kubahlinzeki bangaphandle.

UHaystack, ozothola ku- github.com/deepset-ai/haystack, iwumtapo wolwazi ojwayelekile wokwakha izinjini zokusesha kanye nezinhlelo zemibuzo nezimpendulo ngokusekelwe kumadokhumenti angaphakathi kanye nemithombo yedatha. Ivumela ukuhlanganisa izizindalwazi ezahlukene, izixhumi, kanye namamodeli olimi ukudala izinjini zokusesha zezinkampani ezithuthukisiwe.

Lezi zinjini zokusesha zihlangana kahle kakhulu ne- izinsizakalo zobuhlakani bebhizinisi kanye ne-Power BIngoba zenza kube lula kunoma yisiphi isisebenzi ukuthola ulwazi esiludingayo, futhi ngesikhathi esifanayo, zivumela abahlaziyi ukuthi basebenzise lokho okuqukethwe ukuthola izilinganiso namadeshibhodi. Futhi, ungqimba lwe izicelo ngokwezifiso Yilona elibumba ikhambi lokugcina, ngokungena ngemvume kwezinkampani, ukuhlolwa kanye nokusetshenziswa kwezinsizakalo zamafu ze-AWS kanye ne-Azure uma kudingeka.

I-Stable Diffusion WebUI AUTOMATIC1111: Izinhlelo zokusebenza zokukhiqiza izithombe ze-Clone

Uma unentshisekelo kuzinhlelo zokusebenza ezikhiqiza izithombe ngombhalo (njenge-DALL·E, Midjourney, noma ezifanayo), iphrojekthi okudingeka uyihlole yile I-Stable Diffusion WebUI AUTOMATIC1111Kuyisixhumi esibonakalayo sewebhu esiphelele kakhulu ongasisebenzisa Ukusabalalisa Okuzinzile ngezilungiselelo eziningi kanye nezandiso.

Indawo yokugcina impahla itholakala ku- github.com/AUTOMATIC1111/stable-diffusion-webui futhi inikeza nezinketho ze ukufakwa ngokuchofoza okukodwa kumishini eminingi. Uma isiqalile ukusebenza, ikuvumela ukuthi udale imifanekiso, izinguqulo zomkhiqizo, izizinda, ukuhlukahluka kwamalogo, noma okuqukethwe kokumaketha ngendlela eguquguqukayo kakhulu.

Ngalesi sisekelo ungakwazi phinda ingqikithi yezinhlelo zokusebenza eziningi zokukhiqiza ubuciko ezisebenzisa i-AI yezentengiselwanoKodwa igcinwe kwingqalasizinda yakho, ngamamodeli akho nezilungiselelo zokuphepha. Izinkampani zokuklama, ukumaketha, noma imikhiqizo zingayisebenzisa ukwenza imikhankaso yokubonisa, ukudala imikhiqizo esetshenzisiwe futhi ulungiselele izinto ezibonakalayo ozozethula kumakhasimende.

Ukungenela ikhambi lochwepheshe kuhilela ukwengeza ungqimba lwe isofthiwe yangokwezifiso Kufanele ikhawulele ukuthi ubani ongayifinyelela, abhale ukusetshenziswa, aphathe imigqa yokukhiqiza izithombe, futhi ahlanganiswe nezinhlelo zamafu (isibonelo, ama-GPU ku-AWS noma ku-Azure). Ngale ndlela, ungaba nohlelo lokusebenza lobuciko "lwangaphakathi" olunamandla afanayo namathuluzi amaningi akhokhelwayo.

Ukuphindaphinda Izwi Ngesikhathi Sangempela: ukuphindaphinda amazwi ngendlela efanele

Olunye uhlobo lohlelo lokusebenza oluhlaba umxhwele kakhulu yilolo oluthi ukwenziwa kwezwi kusukela kumasekhondi ambalwa nje omsindo. Iphrojekthi Ukwenziwa Kwezwi Ngesikhathi Sangempela, isingathwe github.com/CorentinJ/Real-Time-Voice-CloningIkuvumela ukuthi uqeqeshe imodeli ngamasampula amafushane futhi ukhiqize izinkulumo ezizwakala zifana kakhulu nezwi lokuqala.

Ngokombono wobuchwepheshe, lokhu kwenza kube nokwenzeka "ukukopisha" amakhono avela ekuqopheni izwi kwezentengiselwano kanye nezinhlelo zokusebenza zokuhlanganisa ezithuthukisiwe. Kodwa-ke, kubalulekile ukusebenzisa lobu buchwepheshe ngendlela... ukuziphatha okuhle nokuhlonipha imithethonqubongoba ingozi yokuxhashazwa iyabonakala uma uhlaka olucacile lokusetshenziswa lungasungulwa.

Amacala okusetshenziswa okunesibopho afaka phakathi Abasizi bezwi abenziwe ngokwezifiso, izixwayiso zezwi ezenzakalelayo zezinhlelo zesevisi yamakhasimendeLokhu kuhlanganisa ukudala imiyalezo yokubamba noma iziqondiso ezisebenzisanayo, njalo ngemvume ecacile yabantu abasebenzisa amazwi abo. Kuwusizo futhi kumaphrojekthi okufinyeleleka noma okuqeqesha.

Ukufaka into efana nale ekukhiqizeni kudinga umsebenzi onzima ukuphepha kwe-inthanethi kanye nokuphathwa kwedathaukuze uhlelo lusetshenziswe kuphela ezimweni ezivunyiwe futhi kugcinwe irekhodi lamazwi, imibhalo, kanye nemiphumela ekhiqizwayo. Ezandleni zethimba lochwepheshe ku izinhlelo zokusebenza ezenziwe ngokwezifiso kanye ne-AI yamabhizinisi, kungaba yingxenye ebalulekile yesipiliyoni sengxoxo esiyinkimbinkimbi.

I-OpenDevin: Ukukopisha abasizi bokuhlela abasebenzisa i-AI

Amathuluzi asiza ngokuhlela usebenzisa i-AI angenye ingxenye echumayo. I-OpenDevin Kuyiphrojekthi yomthombo ovulekile esebenza njenge umsizi wohlelo: ithola imiyalelo ngolimi lwemvelo futhi ikhiqize ikhodi, izikripthi noma izinguquko ezindaweni zokugcina ezikhona.

Ikhodi iyatholakala ku- github.com/OpenDevin/OpenDevin futhi ivumela ukusetha indawo lapho i-AI ibambisana khona nabathuthukisi abangabantu ukusheshisa intuthuko isofthiwe yangokwezifisoUngaphakamisa izici, izivivinyo ezenzakalelayo, imibhalo, noma amathempulethi aphelele ephrojekthi.

Nge-OpenDevin kungenzeka "ukukopisha" lokho okwenziwa amapulatifomu amaningi ekhodi yezentengiselwano, kodwa kuhambisane ne- ubuchwepheshe bangaphakathi kanye nezindinganiso yenkampani yakho. Ungayilungisa ukuze yazi izindawo zakho zokugcina, imitapo yolwazi, kanye neziqondiso zesitayela, ukuze iziphakamiso zivumelane nesimo sakho.

Ezingeni lebhizinisi, lokhu kusho ukuthi amaphrojekthi okuthuthukisa izinhlelo zokusebenza ezenziwe ngokwezifiso ngokusheshangamaphutha ambalwa aphindaphindayo kanye nemibhalo engcono. Iphinde ivule umnyango wokusebenzisa ama-ejenti e-AI azenzakalela imisebenzi ephindaphindayo, njengokukhiqiza ama-skeleton e-API, ukufuduka kwekhodi, noma izikripthi zokuthunyelwa kumasevisi e-AWS kanye ne-Azure cloud.

ULeon: Abasizi bomuntu siqu abavulekile bokukhiqiza

Uma ufuna "ukukopisha" ulwazi lokuba ne- umsizi wezwi lomuntu siqu Njengo-Alexa, uSiri, noma u-Google Assistant, kodwa ngokulawula okuphelele kwepulatifomu, iphrojekthi yakho iyasebenza. LeonLo msizi womthombo ovulekile utholakala ku- github.com/leon-ai/leon futhi iklanyelwe ukusebenza endaweni, ngamamojula angenziwa ngezifiso.

ULeon ulawulwa ngu izwi noma umbhalo Futhi ingaphatha imisebenzi efana nezikhumbuzi, imibuzo esheshayo, ukuhlanganiswa nezinsizakalo zewebhu, noma ukwenziwa kwesikripthi endaweni yakho. Inzuzo enkulu ukuthi ungakwazi dala amamojula akho ukuze ivumelane nezidingo ezithile zenkampani yakho.

Izixazululo eziningi zokukhiqiza zangaphakathi zingakhiwa phezu kukaLeon, zihlanganiswe no Izinsizakalo zefu ze-AWS ne-Azure ukuvumelanisa amakhalenda, ukuphatha imisebenzi, noma ukuthola izaziso ezivela ezinhlelweni zezinkampani. Njengoba ungumthombo ovulekile, unenkululeko yokushintsha indlela oziphatha ngayo, isikhombikubona, kanye nokugeleza kwedatha.

Ngokusekelwa okufanele okuvela eqenjini elikhethekile e ubuhlakani bokwenziwa, izinhlelo zokusebenza ezenziwe ngokwezifiso kanye nokuphepha kwe-inthanethiULeon angaba yindawo yokufinyelela ejwayelekile yabasebenzi bakho: umsizi oqonda umongo webhizinisi lakho, ohlonipha izinqubomgomo zakho, futhi ongathunyelwa noma kuphi lapho unquma khona.

Igalari ye-Google AI Edge: Amarobhothi endawo kuselula ukuze akopishe izinhlelo zokusebenza ze-AI

Ngaphandle kwamaphrojekthi edeskithophu noma eseva, ungakopisha futhi izinhlelo zokusebenza eziningi zeselula ze-AI usebenzisa Google AI Edge Gallery, uhlelo lokusebenza lwe-Android oluvulekile olwenzelwe ukusebenza amamodeli obuhlakani bokwenziwa endawo kudivayisi ngokwayo.

Nale lokusebenza unga bamba izingxoxo nama-bot, udale izithombe, noma unikeze izeluleko zokuklamaKonke lokhu kwenziwa ngaphandle kokuthi idatha iphume efonini yakho. Indlela icacile: konke ukucubungula kwenziwa endaweni, okunikeza ubumfihlo obengeziwe uma kuqhathaniswa nezinsizakalo zamafu. Ngokubuyisela, amandla okucubungula ala mamodeli ngokuvamile angaphansi kwalawo ezixazululo ezifana ne-ChatGPT, i-Copilot, i-Gemini, noma i-DeepSeek.

I-Google AI Edge Gallery uhlelo lokusebenza umthombo ovulekile futhi ovulekileYakhelwe abasebenzisi abanolwazi oluthile. Ayilandwanga ku-Google Play; kufanele uthole i-APK yayo ku-[source missing]. github.com/google-ai-edgeesigabeni Qalisa Emaminithini, lapho uzobona khona isixhumanisi esiya ku- i-APK yakamuvaNgemuva kwalokho kuzodingeka uvumele ukufakwa kwezinhlelo zokusebenza ezivela emithonjeni engaziwa kudivayisi yakho yeselula ukuze uzisebenzise.

Ngaphakathi kohlelo lokusebenza kukhona izigaba ezintathu eziyinhlokoIngxenye eyodwa igxile ekudalweni kwezithombe, enye ekuklanyweni okusheshayo, kanye nengxenye yesithathu yokuxoxa ne-AI, okufana kakhulu nalokho okunikezwa yi-ChatGPT. Ingxenye ngayinye ibonisa amamodeli atholakalayo, usayizi wawo, kanye nenketho yokuwalanda nokuwafaka efonini yakho. Ungangenisa ngisho namafayela emodeli ye-AI osuvele unawo efonini yakho usebenzisa inkinobho ethi + ekhoneni elingezansi kwesokudla.

Le ndlela ivulela abathuthukisi kanye nezinkampani ithuba lokuthi i-prototype AI ihlangabezana ngqo kudivayisiIkopisha imisebenzi eminingi evela kuzinhlelo zokusebenza zezentengiselwano, kodwa ngokuguquguquka kokusebenzisa amamodeli omthombo ovulekile futhi ngaphandle kokuthumela idatha kubantu besithathu. Iyisisekelo esifanele samaphrojekthi ku izinhlelo zokusebenza zeselula ezenziwe ngokwezifisoikakhulukazi uma kuhlanganiswa nama-backend avikelekile asetshenziswa kumasevisi efu le-AWS ne-Azure.

I-Base44: ipulatifomu eguquguqukayo yokwenza ukugeleza kwebhizinisi kube yi-cloning

Ngaphezu kwamaphrojekthi ahlobene ngqo nolimi, umbono, noma amamodeli ezwi, kunezinkundla ezifana nalezi Isisekelo44 ezigxile ekwakheni izicelo zebhizinisi ngokusheshaNakuba kungeyona imodeli ye-AI kuphela, ihlangana kahle kakhulu nezinsizakalo ezihlakaniphile futhi ikuvumela ukuthi "ukopishe" imisebenzi eminingi ejwayelekile evela kumathuluzi okuphatha.

Nge-Base44 ungakha izinhlelo zokusebenza zokukhiqiza zomuntu siqu, amathuluzi asehhovisi elingemuva, amaphothali amakhasimende noma izixazululo zokwenza ngokuzenzakalela kwenqubo yebhizinisi. Futhi kuyindlela enhle kakhulu yokusetha ama-prototypes kanye nama-MVP, ukuhlola imibono yomkhiqizo ngaphandle kwesidingo sokuthuthukisa yonke ingqalasizinda kusukela ekuqaleni.

Amandla angempela avela uma uhlanganisa le nkundla namaphrojekthi e-AI esiwabonile: ungakha i-portal yamakhasimende nge I-chatbot yangaphakathi isekelwe ku-GPT4All noma ku-llama.cpp, faka injini yokusesha neHaystack, engeza imojula yokubhala nge-Whisper noma uhlanganise ijeneretha yesithombe usebenzisa i-Stable Diffusion.

Kumabhizinisi, lokhu kuholela ekusheshiseni ukudalwa kwe isofthiwe yangokwezifiso eliphinda imisebenzi yezixazululo ezidumile ze-niche, kodwa elivumelaniswe nemithetho yebhizinisi lakho, izinhlelo zakho ezikhona, kanye nezidingo zakho ukuphepha kwe-inthanethi kanye nobuhlakani bebhizinisiUmphumela: ukulawula okwengeziwe, ukuncika okuncane kumalayisense, kanye nezixazululo ezihambisana kahle nezinqubo zakho.

Indlela yokukhetha iphrojekthi efanele isimo sakho

Njengoba kunezinketho eziningi kangaka zomthombo ovulekile ezitholakalayo, umbuzo omkhulu uwukuthi Indlela yokukhetha iphrojekthi ye-AI efanelana kahle kakhulu ngalokho ofuna "ukukopisha" noma ukukopisha. Impendulo itholakala ekuchazeni inhloso yebhizinisi kuqala, bese, kusukela lapho, ukuyihlanganisa namathuluzi afanele kakhulu.

Uma udinga bhala umsindo ube umbhalo Ukuze kukhiqizwe imizuzu noma izinhlelo zokusesha eziphakelwayo, i-Whisper iyisinqumo semvelo. ama-agent azimele Kulabo abanomsebenzi wokuhlela nokwenza imisebenzi, i-AutoGPT ilungele kangcono. Ukwakha ama-chatbot endawoInketho engcono kakhulu ukuhlola i-GPT4All noma i-llama.cpp ngokusekelwe ku-hardware yakho kanye nemikhawulo yobumfihlo.

Uma kugxilwe ekubuzeni imibuzo mayelana nemibhalo yangaphakathi, indlela ekahle ukucabangela I-PrivateGPT noma i-HaystackKuye ngokuthi ufuna into evaliwe kakhulu futhi yendawo noma ipulatifomu eguquguqukayo kakhulu ehlanganisa imithombo eminingi. Ngokuqondene nesici sokubuka, i-Stable Diffusion WebUI izokuvumela ukuthi ulingise izinhlelo zokusebenza eziningi zokwenza izithombe, kuyilapho i-Real Time Voice Cloning inikeza isisekelo sezixazululo ze izwi lokwenziwa okwenziwe ngokwezifiso kakhulu.

Uma ufuna ukuthuthukisa umkhiqizo wamaqembu akho okuthuthukisa, I-OpenDevin Kuyindawo enhle yokuqala, futhi yokunikeza abasebenzi noma abasebenzisi umsizi womuntu siqu olawulwa yizwiULeon angaba umngane wakho omkhulu. Kuselula, i-Google AI Edge Gallery ivula umnyango wolwazi lwe-AI lwendawo ku-Android, futhi amapulatifomu afana ne-Base44 enza kube lula ukuhlanganisa zonke lezi zingcezu ndawonye. izicelo zenkampani ezihambisanayo.

Uma umgomo uwukushintsha kusuka ekuhlolweni kwangaphakathi uye ezixazululweni zebhizinisi eziqinile, kunengqondo kakhulu ukuthembela ezinkampanini ezingochwepheshe ukuthuthukiswa kwesofthiwe ngokwezifiso, ubuhlakani bokwenziwa, ukuphepha kwe-inthanethi, kanye nezinsizakalo zamafu ze-AWS kanye ne-Azure, ekwazi ukudala izakhiwo ezikwaziyo ukukhuliswa, eziphephile ezihlanganiswe kahle nesimiso sakho samanje sezinto eziphilayo.

Kungani uthembele kumlingani ochwepheshe ukuze ahlanganise la maphrojekthi?

Nakuba wonke amaphrojekthi esiwabonile evulekile futhi atholakala ku-GitHub, ukuwasebenzisa ekukhiqizeni ngaphakathi kwenkampani akulula njengokwenza ikhosombe bese ucindezela inkinobho. Kudinga... Isipiliyoni ekwakhiweni kwesofthiwe, ukuthunyelwa kwamafu, ukuphepha kwe-inthanethi, kanye nokuphathwa kwedatha.

Umlingani okhethekile ofana I-Q2BSTUDIO, nomlando omuhle ku- Izinhlelo zokusebenza ezenziwe ngokwezifiso, isofthiwe eyenziwe ngokwezifiso, ubuhlakani bokwenziwa kanye ne-AI yamabhizinisiIngaphatha ukuhlaziya izinqubo zakho, ukukhetha izingxenye ze-AI ezinengqondo kakhulu, nokuklama ikhambi elivela ekugcineni kuya ekugcineni: kusukela ku-interface yomsebenzisi kuya kumasevisi efu.

Lolu hlobo lweqembu luqinisekisa ukuthi ikhambi lihambisana nemithethonqubo yomkhakha wakho nokuthi kusetshenziswa imikhuba emihle kakhulu. ukuphepha kwe-cyberukuthi ukuhlanganiswa nezinhlelo zakho zasehhovisi elingemuva kuzinzile nokuthi ingqalasizinda e- Izinsizakalo zefu ze-AWS ne-Azure ilungiselelwe ngokwezindleko kanye nokusebenza kahle.

Ngaphezu kwalokho, zingaphelelisa amaphrojekthi omthombo ovulekile nge izinsizakalo zobuhlakani bebhizinisi kanye ne-Power BIukudala amadeshibhodi abonisa umthelela wangempela we-AI ezinqubweni zakho: izikhathi ezincishisiwe, ikhwalithi ethuthukisiwe, ukwaneliseka kwamakhasimende okwandisiwe, njll. Ngale ndlela awugcini nje "ngokukopisha" ukusebenza kusuka kuzinhlelo zokusebenza ezikhona, kodwa ufinyelela inzuzo yokuncintisana ehlukile.

Ngamafuphi, uhlelo lwe-AI oluvulekile lubeka amathuluzi anamandla kakhulu etafuleni ukuze kopisha futhi uthuthukise izinhlelo zokusebenza ngobuhlakani bokwenziwaKodwa ukuba nomlingani ochwepheshe yikho okwenza umehluko phakathi kokuhlolwa kanye nesisombululo sochwepheshe esilungele ukukhuliswa.

Isamba samaphrojekthi afana ne-Whisper, AutoGPT, GPT4All, PrivateGPT, Stable Diffusion WebUI, Haystack, Real Time Voice Cloning, OpenDevin, Leon, Google AI Edge Gallery, kanye namapulatifomu afana ne-Base44 abonisa ukuthi namuhla kungenzeka ngokuphelele ukuphinda, ukuhlanganisa, nokwedlula imisebenzi eminingi yezinhlelo zokusebenza ezinkulu ze-AI ngokusebenzisa isofthiwe eyenziwe ngokwezifiso, izinsizakalo zamafu, ukuphepha kwe-inthanethi okuqinile, kanye nobuhlakani bebhizinisiUkwakha izixazululo zethu ezihlonipha ubumfihlo, ezihambisana nezinhlelo zakho, futhi eziklanyelwe ngqo indlela inhlangano yakho esebenza ngayo.

I-athikili ehlobene:

I-ChatGPT ku-Mac! Funda indlela yokuxhuma i-AI nezinhlelo zakho zokusebenza eziyintandokazi

Isaka

Umbhali oshisekayo ngomhlaba wamabhayithi nobuchwepheshe ngokujwayelekile. Ngiyathanda ukwabelana ngolwazi lwami ngokubhala, futhi yilokho engizokwenza kule bhulogi, ngikubonise zonke izinto ezithakazelisayo kakhulu ngamagajethi, isofthiwe, ihadiwe, izitayela zobuchwepheshe, nokuningi. Inhloso yami ukukusiza ukuthi uzulazule emhlabeni wedijithali ngendlela elula nejabulisayo.