scripts/convert_voice.py

#!/usr/bin/env python3
"""Convert a Kokoro PyTorch voice tensor to a portable raw float32 matrix."""

from __future__ import annotations

import argparse
from pathlib import Path

import numpy as np
import torch


def main() -> None:
    parser = argparse.ArgumentParser()
    parser.add_argument("source", type=Path)
    parser.add_argument("destination", type=Path)
    args = parser.parse_args()

    voice = torch.load(args.source, map_location="cpu", weights_only=True)
    values = voice.detach().cpu().numpy().astype("<f4", copy=False).reshape(510, 256)
    if values.shape != (510, 256) or not np.isfinite(values).all():
        raise RuntimeError(f"Invalid Kokoro voice tensor: {args.source}")
    args.destination.parent.mkdir(parents=True, exist_ok=True)
    values.tofile(args.destination)
    print(args.destination)


if __name__ == "__main__":
    main()