Foundational Models for State-of-the-Art Speech and Text Translation