---
dataset_info:
  features:
  - name: ground_truth
    dtype: string
  - name: pixel_values
    sequence:
      sequence:
        sequence: float32
  - name: labels
    sequence: int64
  - name: target_sequence
    dtype: string
  splits:
  - name: train
    num_bytes: 13289197791.6
    num_examples: 900
  - name: test
    num_bytes: 1476577532.4
    num_examples: 100
  download_size: 479927412
  dataset_size: 14765775324.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---