README.md

metadata

title: MarmaSpeakTTS Demo
emoji: 🗣️
colorFrom: indigo
colorTo: blue
sdk: gradio
sdk_version: 5.23.3
app_file: app.py
pinned: false

MarmaSpeakTTS Demo

This Hugging Face Space demonstrates the MarmaSpeakTTS model (CLEAR-Global/marmaspeak-tts-v1), a text-to-speech synthesis model for the Marma language (ISO code: rmz).

About the Model

MarmaSpeakTTS is a text-to-speech model for the Marma language (also called Rakhine), a Tibeto-Burman language spoken by the Marma people in Bangladesh and Myanmar. The model was fine-tuned from Facebook's Massively Multilingual Speech (MMS) VITS model.

Usage

Enter Marma text in the input box or select a sample sentence Click "Synthesize" or "Use Selected Sample" Listen to the generated speech

Technical Details

Base model: Massively Multilingual Speech (MMS) Fine-tuned for Marma language Training data: Marma language audio recordings collected by CLEAR Global Repository: CLEAR-Global/marmaspeak-tts-v1

License

MIT