README.md

---
dataset_info:
  features:
  - name: key
    dtype: string
  - name: ocr
    struct:
    - name: images_bbox
      sequence:
        sequence: float64
    - name: images_bbox_no_text_overlap
      sequence:
        sequence: float64
    - name: lines
      struct:
      - name: bbox
        sequence:
          sequence: float64
      - name: score
        sequence: float64
      - name: text
        sequence: string
      - name: word_slice
        sequence:
          sequence: int64
    - name: words
      struct:
      - name: bbox
        sequence:
          sequence: float64
      - name: line_pos
        sequence:
          sequence: int64
      - name: score
        sequence: float64
      - name: text
        sequence: string
  - name: lang
    dtype: string
  - name: ground_truth
    struct:
    - name: figures
      list:
      - name: bbox
        struct:
        - name: height
          dtype: float64
        - name: left
          dtype: float64
        - name: top
          dtype: float64
        - name: width
          dtype: float64
      - name: caption
        dtype: string
      - name: description
        dtype: string
    - name: footer
      dtype: string
    - name: forms
      sequence: 'null'
    - name: header
      dtype: string
    - name: page
      dtype: int64
    - name: tables
      sequence: 'null'
    - name: text
      dtype: string
  - name: valid_pydantic_model
    dtype: bool
  - name: text_len
    dtype: int64
  - name: num_columns
    dtype: int64
  - name: hierarchical_depth
    dtype: int64
  - name: incorrect_crlf
    dtype: int64
  - name: image
    dtype: image
  - name: prediction
    dtype: string
  splits:
  - name: train
    num_bytes: 7895967.0
    num_examples: 55
  download_size: 6681065
  dataset_size: 7895967.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---