---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: annotation
    struct:
    - name: document
      struct:
      - name: image
        struct:
        - name: height
          dtype: int64
        - name: width
          dtype: int64
      - name: language
        dtype: string
      - name: lines
        list:
        - name: line_id
          dtype: int64
        - name: polygon
          sequence: int64
        - name: reading_order
          dtype: int64
        - name: text
          dtype: string
        - name: words
          list:
          - name: akhar_units
            sequence: string
          - name: polygon
            sequence: int64
          - name: reading_order
            dtype: int64
          - name: text
            dtype: string
          - name: word_id
            dtype: int64
      - name: script
        dtype: string
  splits:
  - name: train
    num_bytes: 269629927.91999996
    num_examples: 1640
  - name: eval
    num_bytes: 78106556.0
    num_examples: 468
  - name: test
    num_bytes: 38356530.0
    num_examples: 236
  download_size: 341408285
  dataset_size: 386093013.91999996
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: eval
    path: data/eval-*
  - split: test
    path: data/test-*
---