---
dataset_info:
  features:
  - name: images
    dtype: image
  - name: texts
    dtype: string
  - name: document_text
    dtype: string
  - name: doc_type
    dtype: string
  - name: page_label
    dtype: string
  - name: render_style
    dtype: string
  - name: task
    dtype: string
  - name: prompt
    dtype: string
  - name: answer
    dtype: string
  - name: messages
    dtype: string
  splits:
  - name: train
    num_bytes: 669823197
    num_examples: 4800
  download_size: 689112507
  dataset_size: 669823197
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---