---
dataset_info:
  features:
  - name: pixel_values
    sequence:
      sequence:
        sequence: float32
  - name: labels
    sequence: int64
  splits:
  - name: train
    num_bytes: 31996610100
    num_examples: 18025
  download_size: 2717059468
  dataset_size: 31996610100
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---