Local LLM

AlexPechkin

Mayelana nalolu hlelo lokusebenza

I-Local LLM iyithuluzi elinamandla, eligxile kubumfihlo elenzelwe ukusebenzisa amamodeli ezilimi ezinkulu ngqo kudivayisi yakho. Akukho ifu, akukho ukuqoqwa kwedatha, futhi akukho ukubhalisa okudingekayo - ukusebenza kwe-AI yendawo kuphela.


Izici Eziyinhloko:

• Amakhono e-Multimodal kanye nokusesha iwebhu: Thumela izithombe nomsindo ngqo ku-AI yakho ngokusekelwa kwamamodeli e-Gemma 4 e-multimodal ngefomethi ye-LiteRT (.littert). Ngaphezu kwalokho, ngokuqinisekiswa kwakho okucacile, lawa mamodeli angasesha iwebhu ukuze alande ulwazi lwesikhathi sangempela.

• Ukusekelwa kwe-GGUF: Ngenisa futhi usebenzise ama-LLM akho owathandayo alungiselelwe imiyalelo ngefomethi ye-GGUF (njenge-Gemma nokuningi).

• Isitolo Semodeli: Phequlula futhi ulande amamodeli anconywe ngunjiniyela ngqo kuhlelo lokusebenza - noma ungenise awakho.

• Ukufakwa Kwezwi nge-Whisper noma i-Gemma 4: Sebenzisa ukuqashelwa kwenkulumo okusekelwe ku-Whisper noma ku-Gemma 4 ukuze uxhumane namamodeli akho ngaphandle kwezandla, ngokugcwele kudivayisi.

• I-TTS yesi-Ukraine kanye ne-Text Reader: Jabulela i-text-to-speech reader ehlanganiswe ngokugcwele enikwe amandla yi-StyleTTS2 (yi-patriotyk), ngokusekelwa kolimi lwesi-Ukraine lwendabuko — akukho i-inthanethi edingekayo.
• Imodi Yomhumushi we-AI: Sebenzisa amamodeli akho endawo njengomhumushi ochwepheshe ngenkathi ugcina okuqukethwe kwakho kuyimfihlo ngokuphelele.

• 100% Akuxhunyiwe Ku-inthanethi Futhi Kuyimfihlo: Idatha yakho ayilokothi ishiye idivayisi yakho. Konke ukucubungula kwenziwa endaweni, kuqinisekisa ubumfihlo nokuphepha okuphezulu.

Kungani ukhetha i-LLM Yendawo?
• Azikho izinkinobho ze-API noma uxhumano lwe-inthanethi oludingekayo ngemuva kokusethwa kwemodeli.

• Ukulawula okugcwele amamodeli akho e-AI kanye nedatha.
• Ilungele onjiniyela, abathanda ubumfihlo, nanoma ubani ofuna ukuhlola umhlaba we-AI yendawo.

Inothi Elibalulekile:
Uhlelo lokusebenza alufaki amafayela emodeli afakwe ngaphambilini. Sebenzisa i-Model Store eyakhelwe ngaphakathi ukulanda amamodeli anconyiwe, noma ungenise amafayela akho e-GGUF alungiselelwe imiyalelo.

Yini entsha

Bug fixes.