Self-supervised Speech Representations Still Struggle with African American Vernacular English

Open in new window