ALEPH WEBETA is an Advanced Text-to-Speech (TTS) Model Training Tool with a Graphical User Interface (GUI). This project aims to simplify the process of training custom TTS models, particularly for low-resource languages like Amharic.
Developed by Robel Adugna, this tool provides a user-friendly interface for data preprocessing, model training, and speech generation using state-of-the-art TTS techniques.
- Data Preprocessing: Easily preprocess audio files with options for normalization, noise reduction, and silence trimming.
- Model Training: Train custom TTS models with configurable parameters, including model type, layers, batch size, learning rate, etc.
- Multi-language Support: Adaptable to various languages, with a focus on Amharic and other low-resource languages.
- Custom Alphabet and Cleaning Rules: Define custom character sets and text cleaning rules for your target language.
- Speech Generation: Generate speech from text using trained models.
- Google Drive Integration: Seamlessly work with datasets and models stored on Google Drive.
-
Clone this repository:
git clone https://github.com/yourusername/aleph-webeta.git cd aleph-webeta -
Install the required dependencies:
pip install -r requirements.txt -
(Optional) If using Google Colab, mount your Google Drive:
from google.colab import drive drive.mount('/content/drive')
-
Launch the Gradio interface:
python TTS_Model.py
-
Use the GUI to:
- Preprocess your audio data
- Configure and train your TTS model
- Generate speech from text using your trained model
Contributions to ALEPH WEBETA are welcome! Please feel free to submit pull requests, create issues or spread the word.
This project is licensed under the MIT License - see the LICENSE file for details.
Robel Adugna - [[email protected]|+251913250168| Ethiopia]
Project Link: [https://github.com/robadugna2/ALEPH_WEBETA-TTSMODEL-GUI)
Made with ❤️ for advancing TTS technology in low-resource languages.