---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: bboxes
    sequence:
      sequence: float64
  - name: labels
    sequence: int64
  - name: words
    sequence: string
  - name: split
    dtype: string
  splits:
  - name: train
    num_bytes: 32034973965.125
    num_examples: 80863
  download_size: 0
  dataset_size: 32034973965.125
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---
# Dataset Card for "doclaynet_processed"

Clean version of [DocLayNet](https://github.com/DS4SD/DocLayNet) ready for finetuning.