TTS_FARMLINGUA-models

WavTokenizer for Farmlingua TTS Pipeline — High-quality neural audio tokenizer (codec) used in the Farmlingua Voice System for African farmers.

This repository contains the WavTokenizer checkpoint + config used as the audio tokenizer in the Farmlingua multilingual TTS system (text-to-text + text-to-speech for farmers).

Model Details

Base Model: novateur/WavTokenizer-large-320-24k-4096
Architecture: WavTokenizer (Vocos backbone + single quantizer)
Sampling Rate: 24kHz
Tokens per second: 75
Use Case: Audio tokenization for TTS in low-resource African languages (Hausa, Yoruba, etc.)

How to Use (Direct Loading)

1. Install dependencies

pip install torch torchaudio
git clone https://github.com/jishengpeng/WavTokenizer.git
cd WavTokenizer
pip install -e .

Downloads last month: -; Downloads are not tracked for this model. How to track