ืžืึธื–ื™ืœืœืึท ื”ืื˜ ื‘ืืคืจื™ื™ื˜ ืึท ื“ืขืจื”ื™ื™ึทื ื˜ื™ืงืŸ ืฆื• ื–ื™ื™ืŸ ืงืึธืžืžืึธืŸ ืงื•ืœ ื“ืึทื˜ืึทืกืขืฅ, ื•ื•ืึธืก ืึทืจื™ื™ึทื ื ืขืžืขืŸ ืคึผืจืึธื•ื ืึทื ืกื™ื™ื™ืฉืึทืŸ ืกืึทืžืคึผืึทืœื– ืคื•ืŸ ืงื™ืžืึทื˜ 200 ืžืขื ื˜ืฉืŸ. ื“ื™ ื“ืึทื˜ืŸ ื–ืขื ืขืŸ ืืจื•ื™ืก ื•ื•ื™ ืฆื™ื‘ื•ืจ ืคืขืœื“ (CC0). ื“ื™ ืคืืจื’ืขืœื™ื™ื’ื˜ ืฉื˜ืขืœื˜ ืงืขื ืขืŸ ื–ื™ื™ืŸ ื’ืขื•ื•ื™ื™ื ื˜ ืื™ืŸ ืžืึทืฉื™ืŸ ืœืขืจื ืขืŸ ืกื™ืกื˜ืขืžืขืŸ ืฆื• ื‘ื•ื™ืขืŸ ืจืขื“ืข ื“ืขืจืงืขื ื•ื ื’ ืื•ืŸ ืกื™ื ื˜ืขื– ืžืึธื“ืขืœืก. ืงืึทืžืคึผืขืจื“ ืฆื• ื“ื™ ืคืจื™ืขืจื“ื™ืงืข ื“ืขืจื”ื™ื™ึทื ื˜ื™ืงืŸ, ื“ื™ ื‘ืึทื ื“ ืคื•ืŸ ืจืขื“ืข ืžืึทื˜ืขืจื™ืึทืœ ืื™ืŸ ื“ืขืจ ื–ืึทืžืœื•ื ื’ ื’ืขื•ื•ืืงืกืŸ ืžื™ื˜ 30% - ืคื•ืŸ 13.9 ืฆื• 18.2 ื˜ื•ื™ื–ื ื˜ ืฉืขื” ืคื•ืŸ ืจื™ื™ื“. ื“ื™ ื ื•ืžืขืจ ืคื•ืŸ ืฉื˜ื™ืฆื˜ ืฉืคึผืจืึทื›ืŸ ืื™ื– ื’ืขื•ื•ืืงืกืŸ ืคื•ืŸ 67 ืฆื• 87.

ื“ืขืจ ื’ืึทื ื’ ืคึฟืึทืจ ื“ื™ ืจื•ืกื™ืฉืข ืฉืคึผืจืึทืš ืงืึธื•ื•ืขืจืก 2452 ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 193 ืฉืขื” ืคื•ืŸ ืจืขื“ืข ืžืึทื˜ืขืจื™ืึทืœ (ืขืก ื–ืขื ืขืŸ ื’ืขื•ื•ืขืŸ 2136 ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 173 ืฉืขื”), ืคึฟืึทืจ ื“ื™ ื‘ืขืœืึธืจื•ืกื™ืฉ ืฉืคึผืจืึทืš - 6160 ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 987 ืฉืขื” (ืขืก ื–ืขื ืขืŸ ื’ืขื•ื•ืขืŸ 3831 ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 356 ืฉืขื”), ืคึฟืึทืจ ื“ื™ ืื•ืงืจื™ื™ื ื™ืฉ ืฉืคึผืจืึทืš - 684 ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 76 ืฉืขื” (ืขืก ื–ืขื ืขืŸ ื’ืขื•ื•ืขืŸ 615 ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 66 ืฉืขื”). ืžืขืจ ื•ื•ื™ 79 ื˜ื•ื™ื–ื ื˜ ืžืขื ื˜ืฉืŸ ื”ืึธื‘ืŸ ืึธื ื˜ื™ื™ืœ ื’ืขื ื•ืžืขืŸ ืื™ืŸ ื“ืขืจ ืฆื•ื’ืจื™ื™ื˜ื•ื ื’ ืคื•ืŸ ืžืึทื˜ืขืจื™ืึทืœืก ืื™ืŸ ืขื ื’ืœื™ืฉ, ื“ื™ืงื˜ื™ื™ื˜ื™ื ื’ 2886 ืฉืขื” ืคื•ืŸ ื‘ืืฉื˜ืขื˜ื™ืงื˜ ืจืขื“ืข (ืขืก ื–ืขื ืขืŸ ื’ืขื•ื•ืขืŸ 75 ื˜ื•ื™ื–ื ื˜ ืคึผืึทืจื˜ื™ืกืึทืคึผืึทื ืฅ ืื•ืŸ 2637 ืฉืขื”).

ื–ืืœ ืื•ื ื“ื– ื“ืขืจืžืึธื ืขืŸ ืื™ืจ ืึทื– ื“ื™ ืงืึธืžืžืึธืŸ ืงื•ืœ ืคึผืจื•ื™ืขืงื˜ ืื™ื– ืึทื™ืžืขื“ ืฆื• ืึธืจื’ืึทื ื™ื–ื™ืจืŸ ืฉืœืึธืก ืึทืจื‘ืขื˜ ืฆื• ืึธื ืงืœื™ื™ึทื‘ืŸ ืึท ื“ืึทื˜ืึทื‘ื™ื™ืก ืคื•ืŸ ืงื•ืœ ืคึผืึทื˜ืขืจื ื– ื•ื•ืึธืก ื ืขืžื˜ ืื™ืŸ ื—ืฉื‘ื•ืŸ ื“ื™ ื“ื™ื™ื•ื•ืขืจืกื™ื˜ื™ ืคื•ืŸ ืงื•ืœื•ืช ืื•ืŸ ืจืขื“ืข ืกื˜ื™ื™ืœื–. ื™ื•ื–ืขืจื– ื–ืขื ืขืŸ ืคืืจื‘ืขื˜ืŸ ืฆื• ืงื•ืœ ืคืจืึทืกืขืก ื’ืขื•ื•ื™ื–ืŸ ืื•ื™ืฃ ื“ืขื ืขืงืจืึทืŸ ืึธื“ืขืจ ืึธืคึผืฉืึทืฆืŸ ื“ื™ ืงื•ื•ืึทืœื™ื˜ืขื˜ ืคื•ืŸ ื“ืึทื˜ืŸ ืฆื•ื’ืขื’ืขื‘ืŸ ื“ื•ืจืš ืื ื“ืขืจืข ื™ื•ื–ืขืจื–. ื“ื™ ืึทืงื™ื•ืžื™ืึทืœื™ื™ื˜ื™ื“ ื“ืึทื˜ืึทื‘ื™ื™ืก ืžื™ื˜ ืจืขืงืึธืจื“ืก ืคื•ืŸ ืคืึทืจืฉื™ื“ืŸ ืคึผืจืึธื•ื ืึทื ืกื™ื™ื™ืฉืึทื ื– ืคื•ืŸ ื˜ื™ืคึผื™ืฉ ืคืจืึทืกืขืก ืคื•ืŸ ืžืขื ื˜ืฉ ืจืขื“ืข ืงืขื ืขืŸ ื–ื™ื™ืŸ ื’ืขื•ื•ื™ื™ื ื˜ ืึธืŸ ืจื™ืกื˜ืจื™ืงืฉืึทื ื– ืื™ืŸ ืžืึทืฉื™ืŸ ืœืขืจื ืขืŸ ืกื™ืกื˜ืขืžืขืŸ ืื•ืŸ ืื™ืŸ ืคืึธืจืฉื•ื ื’ ืคึผืจืึทื“ื–ืฉืขืงืก. ืœื•ื™ื˜ ื“ืขืจ ืžื—ื‘ืจ ืคื•ืŸ ื“ื™ ื•ื•ืึธืกืง ืงืขืกื™ื™ื“ืขืจื“ื™ืง ืจืขื“ืข ื“ืขืจืงืขื ื•ื ื’ ื‘ื™ื‘ืœื™ืึธื˜ืขืง, ื“ื™ ื“ื™ืกืึทื“ื•ื•ืึทื ื˜ื™ื“ื–ืฉื™ื– ืคื•ืŸ ื“ื™ Common Voice ืฉื˜ืขืœืŸ ื–ืขื ืขืŸ ื“ื™ ืื™ื™ืŸ-ืกื™ื™ื“ืึทื“ื ืึทืก ืคื•ืŸ ื“ื™ ืงื•ืœ ืžืึทื˜ืขืจื™ืึทืœ (ื“ื™ ืคึผืจื™ื“ืึทืžืึทื ืึทื ืก ืคื•ืŸ ื–ื›ืจ ืžืขื ื˜ืฉืŸ 20-30 ื™ืืจ ืึทืœื˜, ืื•ืŸ ื“ื™ ืคืขืœืŸ ืคื•ืŸ ืžืึทื˜ืขืจื™ืึทืœ ืžื™ื˜ ื“ื™ ืงื•ืœื•ืช ืคื•ืŸ ืคืจื•ื™ืขืŸ , ืงื™ื ื“ืขืจ ืื•ืŸ ืขืœื˜ืขืจ), ื“ื™ ืคืขืœืŸ ืคื•ืŸ ื•ื•ืขืจื™ืึทื‘ื™ืœื™ื˜ื™ ืื™ืŸ ื“ื™ ื•ื•ืขืจื˜ืขืจื‘ื•ืš (ืจื™ืคึผืขื˜ื™ื˜ื™ืึธืŸ ืคื•ืŸ ื“ื™ ื–ืขืœื‘ืข ืคืจืึทืกืขืก) ืื•ืŸ ื“ื™ ืคืึทืจืฉืคึผืจื™ื™ื˜ื•ื ื’ ืคื•ืŸ ืจืขืงืึธืจื“ื™ื ื’ืก ืื™ืŸ ื“ื™ ื“ื™ืกื˜ืึธืจื˜ื™ื ื’ MP3 ืคึฟืึธืจืžืึทื˜.

ืื™ืŸ ื“ืขืจืฆื•, ืžื™ืจ ืงืขื ืขืŸ ื‘ืึทืžืขืจืงืŸ ื“ื™ ืžืขืœื“ื•ื ื’ ืคื•ืŸ ื“ื™ NVIDIA NeMo 1.6 ื˜ืึธืึธืœืงื™ื˜, ื•ื•ืึธืก ื’ื™ื˜ ืžืึทืฉื™ืŸ ืœืขืจื ืขืŸ ืžืขื˜ื”ืึธื“ืก ืคึฟืึทืจ ืงืจื™ื™ื™ื˜ื™ื ื’ ืจืขื“ืข ื“ืขืจืงืขื ื•ื ื’ ืกื™ืกื˜ืขืžืขืŸ, ืจืขื“ืข ืกื™ื ื˜ืขื– ืื•ืŸ ื ืึทื˜ื™ืจืœืขืš ืฉืคึผืจืึทืš ืคึผืจืึทืกืขืกื™ื ื’. NeMo ื™ื ืงืœื•ื“ื– ื’ืจื™ื™ื˜-ืฆื•-ื ื•ืฆืŸ ื˜ืจื™ื™ื ื“ ืžืึธื“ืขืœืก ืคึฟืึทืจ ืžืึทืฉื™ืŸ ืœืขืจื ืขืŸ ืกื™ืกื˜ืขืžืขืŸ ื‘ืื–ื™ืจื˜ ืื•ื™ืฃ ื“ื™ PyTorch ืคืจื™ื™ืžื•ื•ืขืจืง, ืฆื•ื’ืขื’ืจื™ื™ื˜ ื“ื•ืจืš NVIDIA ื ื™ืฆืŸ Common Voice ืจืขื“ืข ื“ืึทื˜ืŸ ืื•ืŸ ืงืึทื•ื•ืขืจื™ื ื’ ืึท ืคืึทืจืฉื™ื™ื“ื ืงื™ื™ึทื˜ ืคื•ืŸ ืฉืคึผืจืึทื›ืŸ, ืึทืงืกืขื ืฅ ืื•ืŸ ืคืืจืžืขืŸ ืคื•ืŸ ืจื™ื™ื“. ื“ื™ ืžืึธื“ืขืœืก ืงืขืŸ ื–ื™ื™ืŸ ื ื•ืฆื™ืง ืคึฟืึทืจ ืจื™ืกืขืจื˜ืฉืขืจื– ื•ื•ืึธืก ืึทื ื˜ื•ื•ื™ืงืœืขืŸ ืงื•ืœ-ื‘ืื–ื™ืจื˜ ื“ื™ืึทืœืึธื’ ืกื™ืกื˜ืขืžืขืŸ, ื˜ืจืึทื ืกืงืจื™ืคึผืฆื™ืข ืคึผืœืึทื˜ืคืึธืจืžืก ืื•ืŸ ืึธื˜ืึทืžื™ื™ื˜ื™ื“ ืจื•ืคืŸ ืกืขื ื˜ืขืจืก. ืคึฟืึทืจ ื‘ื™ื™ึทืฉืคึผื™ืœ, NVIDIA NeMo ืื™ื– ื’ืขื ื™ืฆื˜ ืื™ืŸ ืึธื˜ืึทืžื™ื™ื˜ื™ื“ ืงื•ืœ ื‘ืึทื“ื™ื ื•ื ื’ืก ืคื•ืŸ MTS ืื•ืŸ Sberbank. ื“ื™ NeMo ืงืึธื“ ืื™ื– ื’ืขืฉืจื™ื‘ืŸ ืื™ืŸ Python ื ื™ืฆืŸ PyTorch ืื•ืŸ ืคื•ื ืื ื“ืขืจื’ืขื˜ื™ื™ืœื˜ ืื•ื ื˜ืขืจ ื“ื™ Apache 2.0 ื“ืขืจืœื•ื™ื‘ืขื ื™ืฉ.

ืžืงื•ืจ: opennet.ru

ืœื™ื™ื’ืŸ ืึท ื‘ืึทืžืขืจืงื•ื ื’