VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Zabaliť text do SSML značiek pre presnú kontrolu:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Značky, ktorým vybraný model rozumie — kliknutím ich umiestnite do textu tam, kde sa vyskytujú:
Tento model číta obyčajný text, takže vnorené značky sa ignorujú.Pre emócie založené na značkách prejdite na expresívny model ako Orpheus alebo Bark.
Definovať vlastné výslovnosti (slovo = výslovnosť):
O nás 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.
Najlepšie pre: General-purpose text-to-speech with natural prosody
Prehľadávať všetky VITS hlasyNa prvý pohľad
- Vývojár
- Jaehyeon Kim et al.
- Licencia
- MIT
- Zvieratá
- free
- Rýchlosť
- fast
- Klonovanie hlasu
- Nie
- Jazyky
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Max. počet znakov
- 2000