VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Kpọchie ngwe gị n'ime SSML táàbụ̀ maka nlekọta ziri ezi:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Táàbụ̀ nke móòdù ahụ a họọrọ na-aghọta - pịa ka ịkpụga otu n'ime ngwe gị ebe ọ na-eme:
Móòdù a na-agụ ngwe nkịtị, yabụ na a na-ewepụta inline táàbụ̀. Maka táàbụ̀-n'okpuru n'émóòdù, gbanwee ka móòdù na-egosi ihe dịka Orpheus mọọbụ Bark.
Ndesịta okwu emeredịkachọrọ:
_N'ihe banyere 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.
Ọkachasị maka: General-purpose text-to-speech with natural prosody
Nlegharịa niile VITS ụdaN'ime nlele
- Ńkwádò
- Jaehyeon Kim et al.
- Ikikere
- MIT
- Tier
- free
- Nhazi
- fast
- Nhazi ụda
- Ọ bụghị
- Asụsụ ndị ahụ
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Ụhara Max
- 2000