VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Amlapio' ch testun mewn tagiau SSML er mwyn cael rheoli cywir:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Tags y deall y model dewisiedig - cliciwch i daflu un i' ch testun lle mae' n digwydd:
Mae'r model yma yn darllen testun plaen, felly anwybyddir tagiau mewnlin. I ddelweddu teimlad yn seiliedig ar dagiau, newidiwch i ddelweddu mynegiant fel Orpheus neu Bark.
Diffinio ynganiad addasiedig (gair = ynganiad):
Am 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.
Gorau ar gyfer: General-purpose text-to-speech with natural prosody
Pori Popeth VITS SaesnegYn syth
- Datblygwr
- Jaehyeon Kim et al.
- Trwydded
- MIT
- o Fawrth
- free
- Cyflymder
- fast
- Clonio llais
- Na
- Iaith:
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Uchafswm nodau
- 2000