Phoneme-Level BERT for Enhanced Prosody of Text-to-Speech with Grapheme Predictions
Paper • 2301.08810 • Published
This is a pretraind PLBERT model used when making a base model for StyleTTS 2 or Stylish TTS
- Phoneme level BERT using the ALBERT architecture but trained on phonemes instead of words
- Taken from StyleTTS 2 repo: https://github.com/yl4579/StyleTTS2
- Paper: https://arxiv.org/abs/2301.08810
- Training Code: https://github.com/yl4579/PL-BERT
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support