VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Wrap test tiegħek fil-tags SSML għall-kontroll preċiż:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Tags li l-mudell magħżul jifhem — ikklikkja biex tqiegħed waħda fit-test tiegħek fejn jiġri:
Dan il-mudell jaqra test sempliċi, għalhekk it-tags inline huma injorati.Għal emozzjoni bbażata fuq it-tag, aqleb għal mudell espressiv bħal Orpheus jew Bark.
Iddefinixxi pronunzji tad-dwana (kelma = pronunzja):
Dwar 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.
L-aħjar għal: General-purpose text-to-speech with natural prosody
Ibbrawżja kollox VITS vuċijietDaqqa t'għajn
- Żviluppatur
- Jaehyeon Kim et al.
- Liċenzja
- MIT
- Annimali
- free
- Veloċità
- fast
- Klonazzjoni tal-vuċi
- Nru
- Lingwi
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Karattri massimi
- 2000