VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Wrap texta í SSML tags fyrir nákvæma stjórn:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Merki sem valið líkan skilur — smelltu til að sleppa einu í textann þinn þar sem það gerist:
Þetta líkan les venjulegan texta, þannig að innlínumerki eru hunsuð. Fyrir merki sem byggja á tilfinningum, skiptu yfir í tjáningarlíkan eins og Orpheus eða Bark.
Skilgreindu sérsniðna framburð (orð = framburð):
Um 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.
Best fyrir: General-purpose text-to-speech with natural prosody
Skoða allt VITS raddirÍ hnotskurn
- Forritari
- Jaehyeon Kim et al.
- Leyfi
- MIT
- Tími
- free
- Hraði
- fast
- Raddklóðun
- Nei
- Tungumál
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Hámarksstafir
- 2000