Ubuyobozi bwa Audio AI

Tacotron 2

Tacotron 2 ni sisitemu iherezo-iherezo-y-inyandiko-y-ijambo kuva kuri Google (2017) ihindura inyandiko yanditse mu buryo bwa mel-spectrogramu, iyo ijwi ry’imitsi rihindura imvugo nkubuzima.

2 min somaIbiherutse kuvugururwa

Incamake

It produced audio rivaling human recordings on key benchmarks.

Kwibira cyane

Tacotron 2 ifite ibice bibiri byingenzi. Ubwa mbere, urukurikirane-rwikurikiranya rwitondewe rwitondewe rusoma inyuguti zinyandiko kandi rugahanura mel-spectrogram ikadiri kumurongo. Kodegisi ihindura inyuguti muburyo bwihishe, uburyo bwitondewe bwokwitondera guhuza inyandiko kumurongo wamajwi, kandi decoder ya autoregressive isohora spekrogramu mugihe 'guhagarika ikimenyetso' yiga mugihe imvugo irangiye. Icya kabiri, amajwi yahinduwe ya WaveNet ahindura iyo mel-spectrogramu muburyo bubi. Mugabanye ikibazo murubu buryo, Tacotron 2 yiga prosody, kuvuga, no kwihuta uhereye kumibare hamwe nubuhanga buke bwubuhanga. Yageze ku manota yibitekerezo yegereye amajwi yabigize umwuga, bituma iba ikirangantego muri synthesis-yumvikana-karemano hamwe nicyitegererezo cya TTS nyuma.

Ubushishozi

Mel-spectrogramu ni intera yubwenge hagati yimiyoboro yombi: iroroshye kandi yoroshye kuburyo bwo kwitondera guhanura, nyamara ikungahaye bihagije kugirango amajwi yongere yubake amajwi menshi. Ibibanza-byitondewe birinda kunanirwa bisanzwe nkamagambo yasubiwemo cyangwa yasimbutse urebye ibice byabanje, hamwe na decoder ya autoregressive hamwe na token yamenyesheje guhagarara ireka icyitegererezo gikora interuro-ndende nziza.

Ingaruka z'Ingamba

Kugera no kugera

Itezimbere kugerwaho binyuze mu kwandukura, kuvuga, no guhuza amajwi.

Igiciro na bije

Amatsinda yibitangazamakuru arashobora kohereza amajwi yihuse hamwe na bije nto.

Umuvuduko n'igipimo

Sisitemu ireba abakiriya irashobora gutunganya imikoranire ivugwa murwego runini.

Kazoza ka Tacotron 2

Igishushanyo cya Tacotron 2 ibyiciro bibiri byashishikarije umuraba wa TTS. Abasimbuye byihuse bidasubirwaho nka FastSpeech 2 bakuyeho decoder ikurikirana kugirango yihute kandi itajegajega, kandi amajwi ya WaveNet ubu asimburwa kuri HiFi-GAN cyangwa moderi yo gukwirakwiza. Umwanya urimo ugana ku ndunduro kugeza ku ndunduro na disikuru nyinshi, zerekana, na zeru-zeru sisitemu yo gukwirakwiza amajwi, ariko Tacotron 2 ikomeje kuba urufatiro rw'imiyoboro ishingiye kuri spekrogramu.

Gushyira mu bikorwa Isi

Guha imbaraga amajwi asanzwe yumvikana muri __AIU_PROTECTED_11 __ 'inyandiko-y-imvugo n'ibicuruzwa.

Gukora ibisobanuro byerekana kubitabo byamajwi na podcasts

Gutanga amajwi kubasoma ecran na software igerwaho

Gukora nkibanze byubushakashatsi nintangarugero yo kwigisha imiyoboro ya TTS

Ingaruka & Kurinda

Gukoresha nabi amajwi no kwigira ibyago byiyongera mugihe uruhushya rubuze.

Ukuri kurashobora kugabanuka hejuru yimvugo, imvugo, cyangwa urusaku rwibidukikije.

Amajwi yubukorikori arashobora kwibeshya kumvugo yukuri nta kirango gisobanutse.

Igishushanyo mbonera

1

Shaka uruhushya rusobanutse rwo gufata amajwi, gukoroniza, no gukoresha.

2

Ikizamini cyiza mubiganiro bitandukanye hamwe nuburyo bwimbere.

3

Sobanura igihe umuntu agomba gusuzuma cyangwa kwemeza ibisubizo.

4

Andika amajwi yubukorikori kandi ugumane inyandiko zerekana kubazwa.

Komeza Ubushakashatsi

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.

Tangira ikibazo

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Ubuyobozi bukurikira

Icyitegererezo

Ibibazo bikunze kubazwa

What is Tacotron 2?

Tacotron 2 ni sisitemu iherezo-iherezo-y-inyandiko-y-ijambo kuva kuri Google (2017) ihindura inyandiko yanditse mu buryo bwa mel-spectrogramu, iyo ijwi ry’imitsi rihindura imvugo nkubuzima. Yakoze amajwi arwanya amajwi yabantu ku bipimo byingenzi.

Ni ubuhe buryo bwo kugereranya Tacotron 2 ihanura uhereye ku nyandiko?

Urusobekerane rwa Tacotron 2 rukurikirana ruvuga mel-spectrogram, iyo ijwi rihinduka amajwi.

Niki gihindura spekrogramu ya Tacotron 2 muburyo bwimikorere?

Tacotron 2 ihuza progaramu ya spekrogramu hamwe na majwi ya WaveNet yahinduwe kugirango itange amajwi yanyuma.

Ni ikihe kibazo kwitondera ahantu bifasha gukumira?

Urebye ibice byabanjirije, kwitondera ahantu bigabanya kunanirwa nkamagambo asubirwamo cyangwa yataye.

Nigute Tacotron 2 yamenya igihe cyo guhagarika gutanga imvugo?

Dododer ya autoregressive ihanura guhagarara, kureka icyitegererezo gikora interuro z'uburebure butandukanye.

Ni ubuhe buryo rusange bwububiko bukoreshwa inyandiko-kuri-ecran ya porogaramu ikoresha?

Tacotron 2 ikoresha kodegisi-decoder ikurikirana-ikurikiranye yerekana ikarita yerekana ikarita yerekana ikarita.