• Python 3.13
• Linux/macOS/Windows
• espeak-ng (for text-to-speech functionality)

Linux:

sudo apt update
sudo apt install espeak-ng

macOS:

brew install espeak-ng

Windows: Download and install from espeak-ng releases

curl -LsSf https://astral.sh/uv/install.sh | sh

# Create virtual environment using uv
uv venv -p 3.13

# Activate virtual environment
# Linux/macOS
source .venv/bin/activate

# Windows
.venv\Scripts\activate

# Install in editable mode (recommended for development)
uv pip install -e .

# Or sync dependencies from lock file
uv sync

1. Create a wikipedia.id folder in the project root:

   mkdir -p wikipedia.id

2. Download the parquet file from HuggingFace:

   • Visit: https://huggingface.co/datasets/wikimedia/wikipedia/tree/main/20231101.id
   • Download all parquet files
   • Save them in the wikipedia.id folder

1. Preprocess the data:

   python preprocessing.py

2. Train the model:

   python train.py