Moshi: a speech-text foundation model for real-time dialogue

Open in new window