NVIDIA Riva na okwu NeMo
NVIDIA Riva bụ SDK nke GPU mebere maka mmepụta okwu AI (ASR, TTS, na ntụgharị), ebe NeMo bụ ngwa ọrụ mepere emepe maka ọzụzụ na imezigharị ụdị dị n'okpuru.
Nchịkọta
Together they let developers build fast, customizable voice applications that run on NVIDIA hardware.
Ime miri emi
NeMo (Modules Neural) bụ ntọala PyTorch mepere emepe nke NVIDIA maka iwulite mkparịta ụka AI. Ọ na-ebufe ụdị a zụrụ azụ maka njirimara okwu akpaaka (ASR), ederede-gaa-okwu (TTS), yana ọrụ asụsụ eke, ahaziri dị ka 'modul neuron' nke enwere ike ijikwa ya, ị nwere ike ịhazigharị na data nke gị. Riva bụ akụkụ mbugharị: ọ na-achịkọta ụdị kachasị mma n'azụ ihe nkesa gRPC na-enuba, na-eji TensorRT na Triton Inference Server iji kụọ obere latency n'ọtụtụ. Otu usoro na-arụ ọrụ na-azụ ma ọ bụ na-emegharị ihe nlereanya na NeMo, na-ebupụ ya na usoro Riva, wee na-ejere ya ozi maka nsụgharị ma ọ bụ nhazi oge. Riva na-akwado nkwanye ugwu site na stampụ ọkwa ọkwa okwu, olu TTS neural, ịkpọ okwu okwu, na ọtụtụ asụsụ, ha niile na-ege ntị ka ha na-agba ọsọ nke ọma na NVIDIA GPUs.
Nghọta nka nka
Ọsọ Riva na-abịa site n'ịchịkọta ụdị ya na TensorRT na ijere ha ozi site na Triton, nke na-eme ka kernels, na-etinye nkenke-nkenke (FP16/INT8), yana batches arịrịọ n'otu oge. Ụdị ASR dị ka Conformer-CTC ma ọ bụ Parakeet iyi ọdịyo na obere chunks ka ị na-edobe ọnọdụ ya, na-emepụta transcript akụkụ n'ime iri puku kwuru iri milliseconds. pipeline TTS jikọtara ụdị ụda olu (dịka, FastPitch) na ụda olu olu (dịka, HiFi-GAN) iji mepụta ụdị ebili mmiri ngwa ngwa karịa ozugbo na otu GPU.
Mmetụta atụmatụ
Nweta na iru
Ọ na-eme ka nnweta ya dịkwuo mma site na ndegharị, ịkọ akụkọ, na ntụgharị olu.
Ọnụ ego na mmefu ego
Ndị otu mgbasa ozi nwere ike ibubata ọdịyo a na-egbu maramara ngwa ngwa site na iji obere mmefu ego.
Ọsọ na ọnụ ọgụgụ
Sistemụ na-eche ihu ndị ahịa nwere ike hazie mkparịta ụka n'ọtụtụ buru ibu.
Ọdịnihu nke NVIDIA Riva na Okwu NeMo
NVIDIA na-akwali Riva na NeMo gaa n'ihu ụdị okwu ntọala ọtụtụ asụsụ na njikọta siri ike na ndị nnọchi anya LLM maka ndị enyemaka olu njedebe na njedebe. Na-atụ anya nhazi ka ukwuu (ịkwalite okwu, olu omenala site na nkeji data), mkpọtụ gburugburu ebe obibi siri ike, na ntinye nke na-agbasa GPU nke etiti data ka ọ bụrụ ngwaọrụ dị ka Jetson. Ka NeMo na-etolite n'akụkụ ụdị mmepụta, ahịrị dị n'etiti njirimara okwu, ntụgharị asụsụ, na echiche mkparịta ụka ga-aga n'ihu na-adaba n'ime pipeline dị n'otu.
Mmejuputa n'ezie n'ụwa
Ndebanye aha n'etiti oku na-eme n'ezie na onye na-ebi ndụ na-enyere aka na-eji akara ngosi ọkwa okwu dee oku ndị ahịa
Iwulite olu TTS ahaziri ahaziri maka onye enyemaka mebere site na imezi FastPitch dị mma na NeMo n'ime awa ole na ole nke ndekọ.
Nkọwa okwu dị ndụ na ntụgharị okwu maka nnọkọ vidiyo ma ọ bụ mmemme nkwanye na NVIDIA GPUs
Idozi ihe nlere ASR nke Conformer nke ọma na ngalaba ahụike akọwapụtara ma ọ bụ okwu iwu site na iji NeMo, wee jee ozi site na Riva
Ihe ize ndụ & okporo ụzọ nche
Iji olu eme ihe na ihe egwu mpụta ga-abawanye mgbe nkwenye na-efu.
Izi ezi nwere ike ịdaba n'ofe ụda olu, olumba ma ọ bụ gburugburu mkpọtụ.
Enwere ike imehie ọdịyo sịntetik dị ka ezigbo okwu na-enweghị akara doro anya.
Map mmejuputa
Nweta nkwenye doro anya maka ijide olu, imechi, na ijigharị.
Nwale ogo n'ofe ndị na-ekwu okwu dị iche iche yana ọnọdụ ndabere.
Kọwaa mgbe mmadụ ga-enyocha ma ọ bụ kwado nsonye.
Deba aha ọdịyo sịntetik ma debe ndekọ ihe ndekọ maka ịza ajụjụ.
Nọgide na-eme nchọpụta
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the NVIDIA Riva and NeMo Speech quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ntuziaka na-esote
Metrics Ogo Okwu PESQ na STOI
Ajụjụ a na-ajụkarị
What is NVIDIA Riva and NeMo Speech?
NVIDIA Riva bụ SDK nke GPU mebere maka mmepụta okwu AI (ASR, TTS, na ntụgharị), ebe NeMo bụ ngwa ọrụ mepere emepe maka ọzụzụ na imezigharị ụdị dị n'okpuru. Ha na-ahapụ ndị mmepe ka ha wuo ngwa olu ngwa ngwa, ahaziri nke na-arụ na ngwaike NVIDIA.
Kedu ihe bụ isi dị n'etiti NeMo na Riva?
NeMo bụ ngwa ọrụ mepere emepe maka iwulite na imezigharị okwu na ụdị asụsụ, ebe ngwugwu Riva ma na-eje ozi ụdị ndị ahụ maka iji mmepụta dị ala.
Kedu teknụzụ NVIDIA abụọ Riva na-adabere na ya iji nweta echiche dị ala?
Riva na-ejikọta ụdị ya na TensorRT maka mkpochapụ GPU kachasị mma ma na-ejere ha ozi site na Triton Inference Server, nke na-ejikwa batching na concurrency dị ike.
N'ime pipeline Riva TTS, kedu ọrụ nke ụda olu dị ka HiFi-GAN?
Ụdị ụda olu dị ka FastPitch na-ebu amụma atụmatụ spectrogram sitere na ederede, yana ụda olu dị ka HiFi-GAN na-atụgharị atụmatụ ndị ahụ ka ọ bụrụ ụda ụda ikpeazụ.
Kedu ihe 'gụgharia' ASR na Riva na-enyere aka?
Ngosipụta gụgharia ọdịyo na obere mkpirisi ma na-ewepụta nsonaazụ n'otu oge n'oge dị nso, kama ichere ka okwu niile gwụ.
Kedu ihe atụ nke nhazi nke NeMo na-enyere aka maka ụdị okwu?
NeMo na-ahapụ ndị mmepe ka ha dezie ụdị a zụrụ azụ na data nke ha, dịka ọmụmaatụ ịmegharị ụdị ASR iji mata okwu ahụike pụrụ iche ma ọ bụ iwu.