VITS TTS
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
Lilitkan teks anda dalam tag SSML untuk kawalan tepat:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Tag model dipilih memahami - klik untuk jatuhkan satu ke dalam teks anda di mana ia berlaku:
Model ini membaca teks biasa, jadi tag dalam baris diabaikan. Untuk emosi berdasar tag, beralih ke model ekspresif seperti Orpheus atau Bark.
Tetapkan sebutan tersendiri (perkataan = sebutan):
Tentang 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.
Terbaik untuk: General-purpose text-to-speech with natural prosody
Layari semua VITS suaraDengan sekejap mata
- Pemaju
- Jaehyeon Kim et al.
- Lesen
- MIT
- Tajuk
- free
- Kelajuan
- fast
- Klon suara
- Tidak
- Bahasa
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- Aksara maksimum
- 2000