Tacotron 2
Tacotron 2 bụ usoro ngwụcha na njedebe ederede-gaa na-okwu sitere na Google (2017) na-atụgharị ederede ederede ozugbo na mel-spectrogram, nke ụda olu na-atụgharị n'ime okwu ndụ.
Nchịkọta
It produced audio rivaling human recordings on key benchmarks.
Ime miri emi
Tacotron 2 nwere akụkụ abụọ bụ isi. Nke mbụ, netwọk usoro-na-usoro nwere nlebara anya na-agụ mkpụrụedemede ederede wee buru amụma mel-spectrogram site na etiti. Ihe ngbanwe na-atụgharị mkpụrụedemede ka ọ bụrụ ihe nnọchianya zoro ezo, usoro nlebara anya na ọnọdụ ọnọdụ na-ahazi ederede na okpokolo agba ọdịyo, yana ihe ngbanwe autoregressive na-ewepụta spectrogram ebe 'nkwụsịtụ token' na-amụta mgbe okwu ahụ kwụsịrị. Nke abụọ, ụda olu WaveNet gbanwetụrụ na-atụgharị mel-spectrogram ahụ ka ọ bụrụ usoro ebili mmiri. Site n'ikewa nsogbu ahụ n'ụzọ dị otú a, Tacotron 2 na-amụta prosody, ịkpọ okwu, na pacing site na data site na iji obere injinịa aka. Ọ nwetara akara echiche pụtara nso na ndekọ ndị ọkachamara, na-eme ka ọ bụrụ akara ngosi na njikọ ụda eke na ihe ndebiri maka TTS akwara.
Nghọta nka nka
mel-spectrogram bụ ihe amamihe dị n'etiti netwọkụ abụọ a: ọ dị kọmpat ma dị mfe maka ụdị nlebara anya ịkọ amụma, mana ọ bara ụba nke ukwuu maka ụda olu iji wughachi ọdịyo ntụkwasị obi dị elu. Nlebara anya nlebara anya na ọnọdụ ebe na-egbochi ọdịda a na-ahụkarị dịka okwu ugboro ugboro ma ọ bụ wụsara site n'ịtụle nhazi n'oge gara aga, yana ihe ngbanwe autoregressive nwere akara nkwụsị mmụta na-ahapụ ka ihe nlereanya ahụ jikwaa ahịrịokwu ogologo agbanwe agbanwe nke ọma.
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 Tacotton 2
Nhazi ọkwa abụọ nke Tacotron 2 kpaliri TTS akwara akwara. Ndị na-anọchi anya ngwa ngwa na-anaghị akwụ ụgwọ dị ka FastSpeech 2 wepụrụ decoder usoro maka ọsọ na nkwụsi ike, na Vocoder WaveNet na-agbanwekarị maka ụdị HiFi-GAN ma ọ bụ mgbasa ozi. Ubi ahụ na-aga n'ihu n'ụzọ zuru oke na njedebe ruo na njedebe na ọtụtụ ndị na-ekwu okwu, na-ekwupụta, na sistemụ ụda olu efu, mana Tacotron 2 ka bụ ntụaka ntọala maka pipeline dabere na spectrogram.
Mmejuputa n'ezie n'ụwa
Na-eme ka olu na-ada ụda sitere n'okike na ngwaahịa na enyemaka ederede Google__
Na-emepụta akụkọ pụtara ìhè maka akwụkwọ ọdịyo na pọdkastị
Na-enye olu maka ndị na-agụ ihuenyo na ngwa ngwa nnweta
Ije ozi dị ka usoro nyocha na nkuzi nkuzi maka pipeline TTS neural
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 Tacotron 2 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
Nlereanya Prosody
Ajụjụ a na-ajụkarị
What is Tacotron 2?
Tacotron 2 bụ usoro ngwụcha na njedebe ederede-gaa na-okwu sitere na Google (2017) na-atụgharị ederede ederede ozugbo na mel-spectrogram, nke ụda olu na-atụgharị n'ime okwu ndụ. Ọ mepụtara ihe ndekọ mmadụ na-asọrịta mpi na isi njiri mara.
Kedu nnọchi anya etiti Tacotron 2 na-ebu amụma site na ederede?
Netwọk usoro-na-usoro nke Tacotron 2 na-ebu amụma maka mel-spectrogram, nke ụda olu ga-aghọzi ọdịyo.
Kedu ihe na-atụgharị Tacotron 2's spectrogram ka ọ bụrụ ụdị ebili mmiri n'ezie?
Tacotron 2 jikọtara amụma spectrogram ya na vocoder WaveNet gbanwetụrụ iji mepụta ọdịyo ikpeazụ.
Kedu nsogbu nlebara anya na ọnọdụ ebe na-enyere aka igbochi?
Site n'ịtụle nhazi ndị gara aga, nlebara anya nke ọnọdụ na-ebelata ọdịda dịka okwu a na-ekwughachi ma ọ bụ tụfuo.
Kedu ka Tacotron 2 si mara mgbe ọ ga-akwụsị iwepụta okwu?
The autoregressive decoder na-ebu amụma nkwụsị nkwụsị, na-ahapụ ka ihe nlereanya ahụ jikwaa ahịrịokwu ogologo dị iche iche.
Kedu ụdị nhazi n'ozuzu ka akụkụ ederede-ga-spectrogram na-eji?
Tacotron 2 na-eji ihe ngbanwe-decoder usoro-na-usoro ihe nlebara anya na mkpụrụedemede eserese na okpokolo agba spectrogram.