VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Wrap qoraalka ku SSML tags si loo hubiyo xakamaynta saxda ah:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Tags qaabka la doortay fahmo — riix si aad u hoos mid ka mid ah qoraalka aad halkaas oo uu ka dhacaa:
Model this akhriyo qoraalka caadiga ah, sidaas inline tags waa la iska indho tiri. For tag-ku salaysan dareenka, u dhaqaaqo si ay u muujiyaan qaabka sida Orpheus ama Bark.
Define custom pronunciations (word = dhawaaqa):
About VITS
VITS — Variational Inference with adversarial learning for end-to-end Text-to-Speech — was introduced by Jaehyeon Kim and collaborators in 2021 and became a foundational architecture for modern neural speech. Rather than the older two-stage pipeline, it synthesizes audio in a single parallel end-to-end pass, pairing a variational autoencoder with normalizing flows and a GAN-style adversarial training process to lift naturalness. At about 25M parameters and trained on ~585 hours, it produces natural prosody at fast inference speeds and supports multiple speakers. It serves as a solid general-purpose, free baseline and underpins many later models such as Piper and MeloTTS.
Ugu Fiican: General-purpose text-to-speech with natural prosody
Taabo oo kala soo bax VITS codEeg
- Soo-saarayaasha
- Jaehyeon Kim et al.
- Liisan
- MIT
- Qiyaamaha
- free
- Xawaaraha
- fast
- Duubista Codka
- Ha
- Afaf
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Noocyada ugu badan
- 2000