Skip to main content

utterance-level representation