Nchịkọta Okwu Ọgwụgwụ ruo na Ọgwụgwụ VITS
VITS bụ ụdị ederede gaa na-okwu nke na-atụgharị ederede ozugbo ka ọ bụrụ ụdị ebili mmiri dị egwu n'otu sistemu a zụrụ azụ, na-amapụ ọkpọkọ ụzọ abụọ a na-emebu.
Nchịkọta
Site na ijikọta inference dị iche iche na ọzụzụ na-emegide ya, ọ na-emepụta okwu dị ịrịba ama, nke na-egosipụta.
Ime miri emi
VITS (Ntuba dị iche iche na mmụta ịta ahụhụ maka ngwụcha ruo ọgwụgwụ Text-to-Speech), nke Kim, Kong na Son webatara na 2021, jikọtara echiche atọ nke sistemu ochie debere iche. Onye na-ahụ maka ọnọdụ dị iche iche autoencoder (VAE) na-amụta nnochite anya okwu, na-eme ka mmiri na-edozi ahụ na-eme ka nkesa ahụ dị nro na-agbanwe nke ọma iji jide nkọwa ụda dị mma, na onye na-akpa ókè ụdị GAN na-akwali ụdị ebili mmiri na-emepụta n'eziokwu. N'ụzọ dị mkpa, VITS na-azụ ụdị ụda olu na ụda olu ọnụ karịa ka usoro abụọ, na-ewepụ ndakọrịta nke na-eweda ogo mgbe a zụrụ modul iche iche. Ọ na-ewebatakwa amụma ogologo oge stochastic, yabụ enwere ike iji otu mkpụrụokwu dị iche iche na-ada ụda oge ọ bụla.
Nghọta nka nka
VITS na-eji Monotonic Alignment Search (MAS) dozie nsogbu nhazigharị ahụ, nke na-achọpụta eserese kacha mma n'etiti akara ederede na okpokolo agba n'oge ọzụzụ na-enweghị ndị na-eme ihe na mpụga. A na-agbakọ azụ VAE site na ọdịyo ahụ n'ezie, ebe a na-emezigharị ọnọdụ ederede tupu oge eruo site na ịhazigharị mmiri na-aga ka ọ kwekọọ. Na ntinye, ị ga-enyocha site na ederede tupu ma decode ozugbo ka ọ bụrụ usoro ifegharị, yabụ enweghị mel-spectrogram dị iche na enweghị vocoder dị iche achọrọ.
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 nchịkọta okwu ngwụcha ruo na ngwụcha VITS
VITS zụlitere ezinụlọ nke ndị ga-anọchi anya nke na-achịkwa TTS mepere emepe. VITS2 mere ka ihe owuwu ahụ dị mfe na ịdị mma ka mma, ebe YourTTS na Coqui XTTS nke a na-ejikarị eme ihe na-agbatị ụzọ iji nweta ụda olu efu na ọtụtụ asụsụ. Na-atụ anya ka a na-arụ ọrụ dị iche iche dị mfe, ezigbo oge na ngwaọrụ, mkpuchi ọtụtụ asụsụ ka mma maka asụsụ ndị dị ala, yana ịchịkwa mmetụta mmetụta uche na ụdị ikwu okwu, ebe ọ bụ na njedebe na njedebe bụ ihe mara mma, ntọala ghọtara nke ọma iji wuo ya.
Mmejuputa n'ezie n'ụwa
Coqui TTS na-ebufe ụdị dabere na VITS nke ndị mmepe na-emezigharị nke ọma iji mechie olu onye na-akọ akụkọ maka akwụkwọ ọdịyo.
Ndị enyemaka olu mepere emepe na ngwa Raspberry Pi-class na-eji ụdị VITS kọmpat maka mmepụta okwu na-anọghị n'ịntanetị zuru oke.
Ngwa mmụta asụsụ na-ewepụta ọmụmaatụ ịkpọ okwu okike site na iji ụdị VITS ọtụtụ asụsụ dị ka YourTTS.
Studioslọ egwuregwu egwuregwu Indie na-emepụta ahịrị mkparịta ụka NPC dị iche iche, na-adabere na amụma oge stochastic maka ụda na-abụghị nke robotic.
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 VITS End-to-End Speech Synthesis 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
Nchịkọta Okwu mmetụta uche
Ajụjụ a na-ajụkarị
Kedu ihe bụ VITS End-to-End Speech Synthesis?
VITS bụ ụdị ederede gaa na-okwu nke na-atụgharị ederede ozugbo ka ọ bụrụ ụdị ebili mmiri dị egwu n'otu sistemu a zụrụ azụ, na-amapụ ọkpọkọ ụzọ abụọ a na-emebu. Site na ijikọta ntule dị iche iche na ọzụzụ mmegide, ọ na-ewepụta okwu ebumpụta ụwa n'ụzọ pụtara ìhè.
Kedu ihe acronym VITS pụtara?
VITS na-anọchi anya Mgbanwe dị iche iche yana mmụta mgbagha maka Ederede-gaa-na-Ọgwụgwụ ruo ọgwụgwụ, na-egosipụta isi ihe atọ ya.
Kedu ihe bụ isi ihe dị iche na ụkpụrụ ụlọ n'etiti VITS na ọkpọkọ TTS ọdịnala?
VITS bụ ngwụcha ruo n'ọgwụgwụ: ọ na-amụta ederede-to-waveform n'otu ụdị, na-ezere ndakọrịta nke ụdị ụda olu dị iche na ụda olu.
Kedu ka VITS si amụta nhazi n'etiti ederede na okpokolo agba?
VITS na-eji Monotonic Alignment Search chọpụta nhazi ederede-ga-frame kacha mma n'ime, na-enweghị mpụta mpụta chọrọ.
Kedu ihe kpatara VITS ji gụnye amụma oge stochastic?
Ihe nleba anya ogologo oge stochastic na-enye ohere ka e jiri usoro okike dị iche iche kwuo otu ahịrịokwu, na-ezere ịdị larịị na nke rọbọt.
Kedu akụrụngwa na-akpali mmepụta VITS gaa na ọdịyo na-ada ụda?
VITS na-eji ọzụzụ adversarial (GAN), ebe onye ịkpa ókè na-ekpe ikpe ma ebili mmiri na-ada ụda n'ezie, na-eme ka mma nghọta dịkwuo mma.