Kyutai Delayed Streams Modeling provides implementations of Speech-To-Text and Text-To-Speech models...