I am trying to build a speech translation using RETURNN toolkit with my native language. what is the possible approach to achieve my goal?
