I-NVIDIA ityala imali eyi-1.5 yezigidi zeerandi kwiprojekthi yeVoice Common Mozilla

I-NVIDIA ityala imali eyi-1.5 yezigidi zeedola kwiprojekthi yeMozilla Common Voice. Umdla kwiinkqubo zokuqaphela intetho uvela kwingqikelelo yokuba kule minyaka ilishumi izayo, iteknoloji yelizwi iya kuba yenye yeendlela eziphambili abantu abanxibelelana ngazo nezixhobo ukusuka kwiikhompyuter kunye neefowuni ukuya kubancedisi bedijithali kunye neekhiski.

Ukusebenza kweenkqubo zezwi kuxhomekeke kakhulu kumthamo kunye neentlobo zedatha yezwi ekhoyo kwiimodeli zokufunda zomshini wokuqeqesha. Itekhnoloji yelizwi yanamhlanje igxile ikakhulu ekuqatshelweni kolwimi lwesiNgesi kwaye ayiquki uluhlu oluninzi lweelwimi, ii-accents, kunye neepateni zokuthetha. Olu tyalo-mali luza kunceda ukukhawulezisa ukukhula kwedatha yelizwi likawonke-wonke, lubandakanye uluntu oluninzi kunye namavolontiya, nokwandisa inani labasebenzi beprojekthi abasisigxina.

Masikukhumbuze ukuba iprojekthi ye-Common Voice ijolise ekuququzeleleni umsebenzi odibeneyo ukuqokelela isiseko sedatha yeepatheni zezwi ezithatha ingqalelo ukuhlukahluka kwamazwi kunye neendlela zokuthetha. Abasebenzisi bayamenywa ukuba bathethe amabinzana elizwi aboniswe kwisikrini okanye bavavanye umgangatho wedatha eyongezwe ngabanye abasebenzisi. Isiseko sedatha esiqokelelweyo esineerekhodi zokubizwa ngeendlela ezahlukeneyo kwamabinzana aqhelekileyo entetho yomntu anokusetyenziswa ngaphandle kwezithintelo kwiinkqubo zokufunda koomatshini nakwiiprojekthi zophando.

Ilizwi eliqhelekileyo iseti okwangoku iquka imizekelo yokubiza amagama abantu abangaphezu kwe-164. Malunga neeyure eziliwaka eziliwaka zedatha yelizwi ziqokelelwe kwiilwimi ezingama-9 ezahlukeneyo. Isethi yolwimi lwesiRashiya ihlanganisa abathathi-nxaxheba be-60 kunye neeyure ze-1412 zezinto zokuthetha, kunye nolwimi lwase-Ukraine - abathathi-nxaxheba be-111 kunye neeyure ze-459. Ukuthelekisa, abantu abangaphezu kwe-30 lamawaka bathathe inxaxheba ekulungiseleleni izixhobo ngesiNgesi, bechaza iiyure ze-66 zentetho eqinisekisiweyo. Iiseti ezicetywayo zingasetyenziswa kwiinkqubo zokufunda zoomatshini ukwakha ukuqondwa kwentetho kunye nemifuziselo yokudibanisa. Idatha ipapashwa njengesizinda soluntu (CC1686).

Ngokombhali wethala leencwadi leVosk eliqhubekayo lokuqatshelwa kwentetho, izinto ezingeloncedo kwiseti yeLizwi eliQhelekileyo licala elinye lezinto zelizwi (ubukhulu becala lamadoda aneminyaka engama-20 ukuya kwengama-30 ubudala, kunye nokungabikho kwezinto ezinamazwi abasetyhini. , abantwana kunye nabantu abadala), ukungabikho kokuguquguquka kwisichazi-magama (ukuphindwaphindwa kwamabinzana afanayo) kunye nokusasazwa kweerekhodi kwifomethi yeMP3 egqwethekileyo.

umthombo: opennet.ru

Yongeza izimvo