Imakandi-Labs's picture
Upload folder using huggingface_hub
67dfc0f verified
---
title: Nigerian TTS Data Preprocessor
emoji: 🎙️
colorFrom: green
colorTo: blue
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
---
# Nigerian TTS Data Preprocessor
Preprocess audio datasets for Nigerian TTS training on **FREE CPU**.
## Purpose
- Downloads Nigerian language datasets (Pidgin, Yoruba, Hausa, Igbo, English)
- Encodes audio with WavTokenizer
- Saves preprocessed data to HuggingFace Hub
- Saves GPU costs by running preprocessing on free CPU!
## Usage
1. Enter your HuggingFace write token
2. Select languages to process
3. Run preprocessing
4. Train on RunPod using the new dataset