---
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