B-K/midi-dataset

--- dataset_info: features: - name: track_id dtype: int64 - name: segment_number dtype: int64 - name: audio_score dtype: float64 - name: composer dtype: large_string - name: difficulty dtype: large_string - name: genre dtype: large_string - name: music_period dtype: large_string - name: form dtype: large_string - name: key_signature dtype: large_string - name: piece_number dtype: float64 - name: performer dtype: large_string - name: opus dtype: float64 - name: midi dtype: binary splits: - name: train num_bytes: 10914239426 num_examples: 820944 download_size: 8004731632 dataset_size: 10914239426 configs: - config_name: default data_files: - split: train path: data/train-* --- # [[aria-midi-v1-pruned-ext](https://huggingface.co/datasets/loubb/aria-midi)] with midi column for midi file