FastSpeech da Mara-autoregressive TTS
FastSpeech yana haifar da gabaɗayan siginar magana a cikin layi ɗaya maimakon firam ɗaya a lokaci guda, yana yin kira cikin sauri da kwanciyar hankali.
Dubawa
It solved the slow, error-prone generation that plagued earlier autoregressive models like Tacotron.
Zurfafa nutsewa
Samfurin TTS na farko na jijiyoyi kamar Tacotron 2 masu jujjuyawa ne: suna tsinkaya kowane firam ɗin sauti akan na baya, wanda yake jinkiri kuma yana da saurin tsallakewa ko maimaita kalmomi lokacin da hankali ya ɓace. FastSpeech, wanda Microsoft da Jami'ar Zhejiang suka gabatar a cikin 2019, yana jujjuya wannan ta hanyar tsinkayar duk firam ɗin lokaci ɗaya. Cibiyar sadarwa ta hanyar ciyarwa ta hanyar Transformer tana ɗaukar saƙonnin wayoyi, a sarari tana yin hasashen tsawon lokacin da kowane sautin waya zai šauki tare da mai sarrafa tsayi, kuma yana faɗaɗa jerin zuwa adadin firam ɗin dama kafin samar da siperogram a cikin fasfo ɗaya. FastSpeech 2 ya inganta akan wannan ta hanyar tsinkayar sauti da kuzari kuma, da kuma ta hanyar horar da maƙasudin tsawon lokaci daga daidaitawar tilastawa maimakon kawar da su daga ƙirar malami mai jinkirin, samar da ƙarin yanayi da magana mai iya sarrafawa.
Fahimtar Fasaha
Dabarar maɓalli shine mai sarrafa tsayi. Saboda rubutu da sauti suna da tsayi daban-daban, FastSpeech yana tsinkayar tsawon lokaci ga kowane sautin waya kuma kawai yana maimaita yanayin ɓoye na wannan wayar wanda sau da yawa don dacewa da tsayin sikirin. Wannan jeri a bayyane yana maye gurbin hankali mara ƙarfi. Ƙirƙirar kowane firam a layi ɗaya yana nufin lokacin ƙididdigewa da ƙyar ya dogara da tsayin jimla, kuma cire madauki na atomatik yana kawar da kurakuran tsalle-tsalle da maimaita kalma.
Dabarun Tasiri
Shiga ku isa
Yana inganta samun dama ta hanyar rubutu, ba da labari, da mu'amalar murya.
Kudin da kasafin kuɗi
Ƙungiyoyin kafofin watsa labaru na iya jigilar sauti mai gogewa cikin sauri tare da ƙaramin kasafin kuɗi.
Gudu da sikelin
Tsarin fuskantar abokin ciniki na iya aiwatar da hulɗar magana a mafi girman ma'auni.
Makomar FastSpeech da TTS mara-autoregressive
Haɗin da ba autoregressive yanzu shine tsoho don samarwa TTS saboda yana da sauri, ƙarfi, kuma mai iya sarrafawa. Tsarukan gaba suna matsawa zuwa mafi kyawun sarrafa kayan aiki, ƙaramin jinkiri don aikace-aikacen kai tsaye, da bambance-bambancen ƙarshen zuwa-ƙarshen waɗanda ke tsallake matsakaicin sikirin gaba ɗaya. Yadawa- da ƙirar tushen-zuciya ba autoregressive suma suna tashi, suna haɗawa da daidaitawar FastSpeech tare da ingantacciyar ƙima, yayin da fayyace filaye da sarrafa tsawon lokaci suna kasancewa masu ƙima don daidaitawa, samfuran murya masu bayyanawa.
Aiwatar da Gaskiyar Duniya
Ka'idodin kewayawa na lokaci-lokaci suna haifar da faɗakarwar murya ta bi-da-bi-da-juya nan take ta yin amfani da daidaitaccen tsarin salon FastSpeech.
Tsarin sabis na abokin ciniki na IVR yana canza rubutu mai ƙarfi zuwa magana a ma'auni ba tare da kurakuran tsalle-tsalle ba.
Masu karanta allo masu isa suna samar da sauri, ingantaccen magana don dogayen takardu akan kayan aiki masu sauƙi.
Kayan aikin abun ciki na murya suna barin masu ƙirƙira su tweak farar magana da ƙimar magana kai tsaye, godiya ga fayyace fage na FastSpeech 2 da masu hasashen kuzari.
Hatsari & Tsare-tsare
Rashin amfani da murya da haɗarin kwaikwaya yana ƙaruwa lokacin da aka rasa izini.
Daidaituwa na iya faɗuwa cikin lafuzza, yaruka, ko mahalli masu hayaniya.
Ana iya kuskuren sauti na roba don ingantacciyar magana ba tare da bayyananniyar lakabi ba.
Taswirar Hanya
Sami tabbataccen izini don ɗaukar murya, cloning, da sake amfani.
Gwajin ingantattun masu magana daban-daban da yanayin baya.
Ƙayyade lokacin da dole ne ɗan adam ya duba ko ya amince da abubuwan da aka fitar.
Yi lakabin sauti na roba da kuma adana bayanan da aka tabbatar don yin lissafi.
Ci gaba da Bincike
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 FastSpeech and Non-Autoregressive TTS 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
Jagora na gaba
Kunkuru TTS Autoregressive Synthesis
Tambayoyin da ake yawan yi
What is FastSpeech and Non-Autoregressive TTS?
FastSpeech yana haifar da gabaɗayan siginar magana a cikin layi ɗaya maimakon firam ɗaya a lokaci guda, yana yin kira cikin sauri da kwanciyar hankali. Ya warware jinkirin, tsararraki masu saurin kuskure waɗanda suka addabi samfuran autoregressive na baya kamar Tacotron.
Menene ma'anar 'mara-autoregressive' a cikin mahallin FastSpeech?
Ba autoregressive yana nufin ana samar da firam ɗin a layi ɗaya a cikin fasfo ɗaya ba, ba sharadi ɗaya-ba-daya akan firam ɗin da suka gabata.
Wanne matsala na ƙirar autoregressive kamar Tacotron 2 ke yin FastSpeech musamman?
Hankalin kai-da-kai na iya yin kuskure, yana haifar da tsallake-tsallake ko maimaita kalmomi, kuma ɓata lokaci yana jinkirin; FastSpeech yana gyara duka biyun.
Menene rawar 'mai sarrafa tsawon' a cikin FastSpeech?
Mai sarrafa tsayi yana faɗaɗa fasalin sautin waya ta tsawon lokacin da aka annabta, yana daidaita gajeriyar jerin rubutu zuwa mafi tsayi.
Menene FastSpeech 2 ya ƙara idan aka kwatanta da ainihin FastSpeech?
FastSpeech 2 yana annabta farar ƙarfi da kuzari kuma yana amfani da lokutan daidaitawar tilastawa maimakon jinkirin malami, haɓaka dabi'a da sarrafawa.
Me yasa lokacin furucin FastSpeech yayi girma da tsayin jimla?
Saboda tsara yana da layi ɗaya, ƙara ƙarin firam ɗin baya ninka matakai na jeri kamar yadda ƙwanƙwasa autoregressive ke yi.