def main(argv: list[str] | None = None) -> None
Extract train clips, train the learned filter, and save the checkpoint.
Source: unitrack/benchmarks/hota/train_learned.py:234
