This experiment aims to develop an open-weight, cost-effective speech model using just a few hundred hours of synthetic training data. After training, the model is converted to be compatible with piper for inference purposes.
Currently, the system only supports US English, though support for additional languages is planned for future releases.
| Filename | Description | Size |
|---|---|---|
2cent-tts-v0.1.0.onnx |
FP32 precision model | 63MB |
This project is licensed under the Apache 2.0 License - see the LICENSE file for details.