🐸 v0.0.13
🐞Bug Fixes
💾 Code updates
SpeakerManagerclass for handling multi-speaker model management and interfacingspeaker.jsonfile.- Enabling multi-speaker models with
ttsandtts-serverendpoints. (👑 @kirianguiller ) - Allow choosing a different
noise scalefor GlowTTS at inference. - Glow-TTS updates to import SC-Glow Models.
- Fixing windows support (👑 @WeberJulian )
🚶♀️ Operational Updates
- Refactoring 🐸 TTS installation and allow selecting different scopes (
all, tf, notebooks)for installation depending on the specific needs.
🏅 Model implementations
🚀 New Pre-Trained Model Releases
- SC-GlowTTS multi-speaker English model from our work https://arxiv.org/abs/2104.05557 (👑 @Edresson )
- HiFiGAN vocoder finetuned for the above model.
- Tacotron DDC Non-Binary English model using Accenture's Sam dataset.
- HiFiGAN vocoder trained for the models above.
Released Models
💡 All the models below are available by tts or tts-server endpoints on CLI as explained here.
Models with ✨️ below are new with this release.
- SC-GlowTTS model is from our latest paper in a collaboration with @Edresson and @mueller91.
- The new non-binary TTS model is trained using the SAM dataset from Accenture Labs. Check out their blog post
| Language | Dataset | Model Name | Model Type | TTS version | Download |
|---|---|---|---|---|---|
| ✨ English (non-binary) | sam (acccenture) | Tacotron2-DDC | tts | 😄 v0.0.13 | 💾 |
| ✨ English (multi-speaker) | VCTK | SC-GlowTTS | tts | 😄 v0.0.13 | 💾 |
| English | LJSpeech | Tacotron-DDC | tts | v0.0.12 | 💾 |
| German | Thorsten-DE | Tacotron-DCA | tts | v0.0.11 | 💾 |
| German | Thorsten-DE | Wavegrad | vocoder | v0.0.11 | 💾 |
| English | LJSpeech | SpeedySpeech | tts | v0.0.10 | 💾 |
| English | EK1 | Tacotron2 | tts | v0.0.10 | 💾 |
| Dutch | MAI | TacotronDDC | tts | v0.0.10 | 💾 |
| Chinese | Baker | TacotronDDC-GST | tts | v0.0.10 | 💾 |
| English | LJSpeech | TacotronDCA | tts | v0.0.9 | 💾 |
| English | LJSpeech | Glow-TTS | tts | v0.0.9 | 💾 |
| Spanish | M-AILabs | TacotronDDC | tts | v0.0.9 | 💾 |
| French | M_AILabs | TacotronDDC | tts | v0.0.9 | 💾 |
| Dutch | MAI | TacotronDDC | tts | v0.0.10 | 💾 |
| ✨ English | sam (accenture) | HiFiGAN | vocoder | 😄 v0.0.13 | 💾 |
| ✨ English | VCTK | HiFiGAN | vocoder | 😄 v0.0.13 | 💾 |
| English | LJSpeech | HiFiGAN | vocoder | v0.0.12 | 💾 |
| English | EK1 | WaveGrad | vocoder | v0.0.10 | 💾 |
| Dutch | MAI | ParallelWaveGAN | vocoder | v0.0.10 | 💾 |
| English | LJSpeech | MB-MelGAN | vocoder | v0.0.9 | 💾 |
| 🌍 Multi-Lang | LibriTTS | FullBand-MelGAN | vocoder | v0.0.9 | 💾 |
| 🌍 Multi-Lang | LibriTTS | WaveGrad | vocoder | v0.0.9 | 💾 |
Update Jun 7 2021: Ruslan (Russian) model has been removed due to the license conflict.