GUIDE IA Audio
NVIDIA Riva ak NeMo
NVIDIA Riva SDK la bu gaaw ci GPU ngir defar kàddu IA (ASR, TTS, ak tekki làkk), fekk NeMo mooy jumtukaayu open-source ngir tàggat ak defar model yi ci suuf.
Déglul tegtal biiIA-nettali · ~2 min
Ci xët wii2 simili jàng
Résumé
Ñu bokk may developpeur yi ñu defar aplikaasioŋu baat yu gaaw, yuñ mëna tànn di dox ci hardware NVIDIA.
Plongeur bu xóot
NeMo (Module Neural) mooy kaadar PyTorch bu NVIDIA ngir tabax IA buy waxtaan. Dafay yónnee ay model yuñ tàggat bu njëkk ngir xàmmee kàddu yi ci saa si (ASR), bind-ci-kaddu (TTS), ak liggéeyi làkk wiñ nàmp, ñu defaree ko ni 'module neuronal' yuñ mëna jëfandikoowaat, nga mën ko gëna defar ci say done. Riva mooy wàllu jëfandikoo gi: dafay ëmb ay model yuñ gëna defar ci ginnàaw serwëru gRPC buy daw, di jëfandikoo TensorRT ak Serwëru Inference Triton ngir am latency bu woyof ci escalier. Liggéey bu ñuy xam mooy tàggat wala méngale ab model ci NeMo, yóbbu ko ci formaa Riva, ba noppi jëfandikoo ko ngir transkripsioŋ wala synthese ci jamono dëgg. Riva dafay jàppale xàmmee streaming ak timestamps ci niveau baat, baat TTS neural, diarization waxkat, ak làkk yu bari, ñoom ñépp ñu ngi aju ci dox bu baax ci NVIDIA GPUs.
Doxal ci àdduna dëgg
Transkripsiyoŋu santaru woote ci jamono dëgg ak ndimbalu ndawu liggéey biy bind wootu kiliyaan yi ci timbre waxtu ci niveau baat
Tabax ay baat TTS yu màrk ngir assistant virtuel ci defar FastPitch ci NeMo ci ay waxtu enregistrement yu néew
Sous-titre ak tekki kàddu ngir ndaje wideo wala xew-xew yuy dawal ci NVIDIA GPUs
Defar ab xeetu Conformer ASR ci vocabulaire medikaal wala yoon bu jëm ci domen bi jëfandikoo NeMo, ba noppi nga joxe ko jaaraleko ci Riva
Ëlëgu NVIDIA Riva ak NeMo
NVIDIA mingi puus Riva ak NeMo ci xeetu kàddu yu gëna yaatu, yu gëna bari làkk, ak lëkkaloo bu gëna dëgër ak ndawu liggéey yu LLM ngir jàppale baat ci njeexte ba ci njeexte. Xaarandil gëna riis ci personaalisasioŋ (baatu boosting, baat yuñ personaaliseer ci ay simili yu done), robustness environnement bu gëna am bruit, ak jëfandikoo giy span GPUs data-center ba ci aparey yu melni Jetson. Bi NeMo di jëm kanam ci wetu xeetu generatif yi, dig wi dox diggante xàmmee kàddu, tekki làkk, ak xalaat ci waxtaan dina wéy di gëna fës ci pipeline yuñ boole ci jamono dëgg.
Teg ko ci jëf
Lu tax mu am solo
- Dugg ak yegg. Dafay gëna yombal jëfandikoo gi jaaraleko ci transkripsioŋ, nettali ak interfaasu baat.
- Njëgg ak budget. Ekipu mejaa yi mën nañu yónnee audio bu leer ci anam wu gëna gaaw te seen xaalis gëna néew.
- Gaawaay ak yaatuwaay. Sistem yiy jàkkarloo ak kiliyaan bi mën nañu def waxtaan ci anam wu gëna yaatu.
Moytul
- Jëfandikoo baat ci anam wu jaarul yoon ak niru ak nit dafay gëna yokk sudee nanguwul.
- Jaar-jaar mën na wàññeeku ci aksan yi, dialect yi wala barab yu bari xumbaay.
- Audio synthetik mën nañu ko jaawale ak wax ju dëggu sudee amul etiket bu leer.
Jéego yi ci topp
- Wutal ndigal bu leer ngir jàpp baat bi, klone ko ak jëfandikoowaat ko.
- Saytu kalite ci kàddukat yu bari ak anam yu bari ci ginaaw.
- Mandargal kañ la nit wara xoolaat wala nangu ay génne.
- Etiketu audio synthetik te nga denc dokimaa ci fimu bawoo ngir mëna lim.
Weyal di banneexu
Bulletin bu amul fayda
Jàngal xibaar AI bu bes
Ñatti jaar-jaari IA yuñ firndeel bis bu nekk ci ayu-bis bi, ñu bind leen ci Àngle bu leer. Gratis ba fàww, amul yëgle.
Benn email bis bu nekk ci ayu-bis bi. Benn klik nga dindi abonemaa bi. Dunu musa jaay wala séddoo sa adres.
Nattal sa bopp
Take the NVIDIA Riva and NeMo Speech quiz
Feedback ci saasi ci bépp tontu, ak sertifikaa buñ mëna séddoo ak ID buñ mëna saytu soo jàllee ci njàng mi.
Jàppale njàngum IA te doo fay. AI Understanding mbootaay la buy def te Yàlla tax — amul yëgle, amul paywall, musul am. Defal ab sarax
Laaj yi ñuy faral di laaj
Luy Riva ak NeMo wax?
NVIDIA Riva SDK la bu gaaw ci GPU ngir defar kàddu IA (ASR, TTS, ak tekki làkk), fekk NeMo mooy jumtukaayu open-source ngir tàggat ak defar model yi ci suuf. Ñoom ñaar ñu ngi may developpeur yi ñu defar aplikaasioŋu baat bu gaaw te ñu mëna tànn di dox ci hardware NVIDIA.
Lan mooy wuutale bu njëkk bi am ci digganté NeMo ak Riva?
NeMo mooy jumtukaayu open-source bi ngir tabax ak defar ay modeli làkk ak kàddu, waaye Riva dafay ëmb model yooyu ngir jëfandikoo ay produit yu néew latency.
Which NVIDIA components does Riva documentation identify for optimized model execution and inference serving?
TensorRT and Triton have distinct execution and serving roles; particular batching and latency behavior depends on the deployed pipeline.
Ci gasoduc Riva TTS, lan mooy wareefu vocoder bu melni HiFi-GAN?
Benn xeetu akustik bu melni FastPitch dafay wax luy waaja am ci spectrogram ci mbind, ba noppi ab vocoder neural bu melni HiFi-GAN dafay soppi màndarga yooyu ci jëmmu onde bu mujj bi ñuy dégg.
Lan mooy 'streaming' ASR ci Riva?
Xamme streaming dafay liggéey audio ci ay pàcc yu ndaw ba noppi génne ay resultaa ci diir bu gàtt, du xaar wax ji yépp jeex.
Lan mooy misaalu personaalisasioŋ bi NeMo mëna def ngir modeli kàddu yi?
NeMo dafay may developpeur yi ñu mëna defar model yiñ njëkka tàggat ci seeni done, lu ci melni méngale model ASR ngir xàmmee terminologie medikaal wala yoon buñ jagleel.
Weyal di jàng
Gid yu jëm ci loolu
Tann nañu yeneen njiit ngir topic bii