| Property | Value |
|---|---|
| License | CC-BY-NC-SA-4.0 |
| Author | mbarnig |
| Supported Languages | Luxembourgish, German, French, English, Portuguese |
| Framework | Coqui-TTS (v0.7.1) |
What is lb-de-fr-en-pt-coqui-vits-tts?
This is a multilingual text-to-speech model built using the VITS architecture and trained on a custom dataset. The model specializes in synthesizing speech in five different languages without relying on phoneme conversion, using a specific character set for direct text-to-speech conversion.
Implementation Details
The model was trained from scratch using the Coqui-TTS multilingual VITS-model recipe. It processes raw text input using a defined character set including standard Latin characters and special characters for various European languages. The training was conducted without phoneme usage, making it more straightforward for direct text processing.
- Custom character set including special characters (ß, à, á, â, ã, ä, ç, è, é, ê, ë, í, î, ï, ó, ô, õ, ö, ù, ú, û, ü)
- Integrated punctuation support (!\'(),-.:;? )
- Direct text-to-speech conversion without phoneme intermediary
Core Capabilities
- Multilingual speech synthesis in five languages
- Support for various European language characters
- Live inference demonstration available
- TensorBoard integration for training visualization
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its support of Luxembourgish alongside other major European languages, and its phoneme-free approach to speech synthesis. The custom dataset and character set make it particularly well-suited for European language processing.
Q: What are the recommended use cases?
The model is ideal for applications requiring multilingual text-to-speech capabilities, particularly in European contexts. It's especially useful for projects involving Luxembourgish, which is less commonly supported in TTS systems.