--- dataset_info: features: - name: audio dtype: audio: sampling_rate: 16000 - name: song_id dtype: string - name: structure dtype: string - name: start_time dtype: float64 - name: end_time dtype: float64 - name: tempos dtype: string - name: keys dtype: string - name: instruments dtype: string - name: genres dtype: string - name: moods dtype: string - name: metadata dtype: string splits: - name: train num_bytes: 797925040.0 num_examples: 831 - name: eval num_bytes: 15363261.0 num_examples: 16 download_size: 801646719 dataset_size: 813288301.0 configs: - config_name: default data_files: - split: train path: data/train-* - split: eval path: data/eval-* ---