--- dataset_info: features: - name: audio_token sequence: sequence: sequence: sequence: int64 - name: audio_duration dtype: float64 - name: text dtype: string - name: align sequence: string splits: - name: train num_bytes: 101071185102 num_examples: 2774765 download_size: 18095604545 dataset_size: 101071185102 configs: - config_name: default data_files: - split: train path: data/train-* ---