The Holistic Evaluation of Audio Representations (HEAR). A benchmark of diverse audio tasks for audio ML researchers.