VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Ulawulo oluchanekileyo:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Ii-tags imodeli ekhethiweyo iqonda - nqakraza ukushiya enye kumbhalo wakho apho isenza khona:
Le modeli ifunda umbhalo oqhelekileyo, ngoko ke i-inline tags ilahleka. Uphawu olusekelwe kwi-emotions, tshintshela kwimodeli ebonisa umbono njenge-Orpheus okanye i-Bark.
Chaza ubeko lwephepha
I-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.
Elungileyo: General-purpose text-to-speech with natural prosody
Khangela konke VITS iilizwiKwingxelo
- Umbhekisi phambili
- Jaehyeon Kim et al.
- Ilayisensi
- MIT
- I-Tier
- free
- Isantya
- fast
- Ukuphinda usebenzise ilizwi
- Akukho nanye
- Iilwimi
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Ubukhulu bamagama
- 2000