I-NVIDIA isanda kwethula umbukiso omusha wobuchwepheshe okulindeleke ukuthi uthuthukise indlela esisebenzisana ngayo nobuhlakani bokwenziwa. Lesi sici esisha, esibizwa ngokuthi “Xoxa Nge-RTX,” siklanyelwe ukusebenza ngaphandle komthungo kuma-Windows RTX PC, sisebenzisa amandla e-NVIDIA RTX GPUs ukuletha ukuzizwisa komuntu siqu nokusebenzayo kwe-chatbot. Lobu buchwepheshe buhlose ukuhlinzeka abasebenzisi ngezimpendulo ezisheshayo, ezivikelekile neziqondile komongo, bedweba kumadokhumenti abo kanye namanothi ukuze baqinisekise ukuxhumana okuyimfihlo nokomuntu siqu.
Enhliziyweni ye-"Chat With RTX" imodeli yolimi ye-GPT eyinkimbinkimbi ekwazi ukuhlanganisa izingxoxo nezidingo ezithile zomsebenzisi. Lena akuyona i-chatbot evamile, kodwa uhlelo oluhlakaniphile olukwazi ukucubungula izinhlobo eziningi zamafayela, okuhlanganisa imibhalo yombhalo, ama-PDF, amadokhumenti e-Word, amafayela e-XML kanye nokulotshwa kwevidiyo kwe-YouTube. Lokhu kusebenza ngezindlela ezihlukahlukene kuvumela i-chatbot ukuthi inikeze usizo oluhlobene kakhulu nokuqukethwe komuntu siqu komsebenzisi.
I-NVIDIA RTX AI PCs Chatbot Demo
Esinye sezici eziyinhloko zedemo yobuchwepheshe entsha ye-NVIDIA ukusetshenziswa kwe-retrieval augmented generation (RAG), okuthuthukisa kakhulu ikhwalithi yezimpendulo ze-chatbot. Ukwengeza, idemo ihlanganisa i-TensorRT-LLM, ithuluzi elikhulu lokuthuthukisa imodeli yolimi, eliqinisekisa ukuthi i-chatbot isebenza kahle kakhulu. Ngokusheshisa kwe-RTX, i-chatbot ayigcini nje ngokunemba kodwa futhi ishesha ngendlela emangalisayo, isebenza ngokuqondile ku-Windows RTX PC yomsebenzisi, ngaphandle kokudinga ukucutshungulwa kwamafu.
Onjiniyela, ikakhulukazi, bangathola okuthi “Xoxa Ne-RTX” kuthakazelisa ngoba yakhela kuphrojekthi yereferensi kanjiniyela we-TensorRT-LLM RAG etholakala ku-GitHub. Lokhu kubanikeza ithuba elibalulekile lokuhlola amamodeli e-AI athuthukile futhi okungenzeka bahlanganise ubuchwepheshe obufanayo kumaphrojekthi abo.
Buka le vidiyo ku-YouTube.
Labo abathanda “Xoxa Nge-RTX” kufanele bahlangabezane nezimfuneko ezithile zesistimu. I-PC yomsebenzisi kufanele ifakwe i-GeForce RTX 30 Series GPU noma imodeli ethuthuke kakhulu, enobuncane obungu-8 GB we-VRAM. Ukwengeza, i-PC kufanele isebenze Windows 10 noma i-11 futhi ibe nezishayeli zakamuva ze-NVIDIA ezifakiwe ukuze kuqinisekiswe ukuhambisana nedemo.
I-NVIDIA ivume inkinga yokufaka eqhubekayo ethi “Xoxa Nge-RTX” futhi ithembise ukuyixazulula kusibuyekezo esizayo. Ngaleso sikhathi, abasebenzisi bayelulekwa ukuthi bafake uhlelo kunkomba ezenzakalelayo ukuze bagweme noma yiziphi izinkinga.
Ukwengeza, i-NVIDIA ikhuthaza abathuthukisi ukuthi basunduze imingcele ye-AI ekhiqizayo ngokuhlela umncintiswano. Abahlanganyeli bayamenywa ukuthi bakhe izinhlelo zokusebenza ezintsha besebenzisa i-NVIDIA RTX GPUs, kanye nethuba lokuwina imiklomelo. Lo mncintiswano awugcini nje ngokukhuthaza ubuciko phakathi komphakathi wonjiniyela, kodwa futhi ugqamisa amandla obuchwepheshe be-NVIDIA ukuthuthukisa izinhlelo zokusebenza ze-AI.
Ukwethulwa kokuthi "Xoxa Nge-RTX" kukhombisa imizamo eqhubekayo ye-NVIDIA yokuthuthukisa ubuchwepheshe be-AI ne-GPU. Ngokugxila ekusebenzeni okuphezulu kwekhompuyutha nobumfihlo bedatha, i-NVIDIA inika amandla amakhono e-AI ayinkimbinkimbi ukuthi ahlanganiswe nemisebenzi yansuku zonke. Lobu buchwepheshe buvumela abasebenzisi ukuthi bahlomule kumsizi we-AI ohlakaniphile, ophendulayo futhi oqondene nawe, kuyilapho begcina idatha yabo ecutshunguliwe ngokuvikelekile emshinini wabo wasendaweni. Njengoba i-NVIDIA iqhubeka nokusungula nokuxazulula izinkinga zokukhula, i-"Chat With RTX" isilungele ukuba yithuluzi elibalulekile lalabo abafuna ulwazi lwekhompuyutha oluhlakaniphile, oluphendulayo.
Funda kabanzi Umhlahlandlela:
- Umthombo ovulekile we-Google Gemma i-AI yenzelwe ukusebenza kuma-NVIDIA GPU
- Konke okusha nge-ChatGPT ne-AI kuleli sonto
- Ungawakha kanjani ama-Chatbots e-AI angokwezifiso ngeGoogle Gemini
- I-LibreChat yamahhala neyomthombo ovulekile yemodeli ye-AI esebenzayo
- I-Ollama ye-Windows isiyatholakala ukusebenzisa ama-LLM endaweni
- I-Google Gemini izoyithuthukisa kanjani i-Google Bard
