VITS ټي ټي اېس
The end-to-end TTS architecture that combines a variational autoencoder, normalizing flows, and adversarial training.
د دقیق کنټرول لپاره په SSML نښانونو خپل متن واچوئ:
<speak><prosody rate="slow">Slow speech</prosody></speak>
توري د ټاکل شوي ماډل پوهیږي - کلیک وکړئ چې ستاسو په متن کې یو راښکته کړئ چیرې چې دا پیښیږي:
دا ماډل لوستل ساده متن، نو inline نښانونه په پام کې نه نيول کيږي. د نښان پر بنسټ احساس، د يو څرګند ماډل لکه Orpheus يا Bark بدل.
دوديزه لوستنه پېژندل (ويې = لوستنه):
په اړه 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.
غوره د: General-purpose text-to-speech with natural prosody
ټول لټول VITS غږونهپه يوه کتنه کې
- جوړوونکی
- Jaehyeon Kim et al.
- منښتليک
- MIT
- :د پاڼې نوم
- free
- چټکتيا
- fast
- غږ کلونول
- نه
- ژبې
- English, German, Spanish, French, Portuguese, Dutch, Finnish, Hungarian, Bulgarian, Japanese, Polish
- ټولوجګه لوښه
- 2000