Convert the ReadingBank dataset (from LayoutReader: Pre-training of Text and Layout for Reading Order Detection) into a Hugging Face Dataset with train/dev/test splits.

References:

• Paper: https://arxiv.org/abs/2108.11591

• Original Dataset: https://mail2sysueducn-my.sharepoint.com/:u:/g/personal/huangyp28_mail2_sysu_edu_cn/Efh3ZWjsA-xFrH2FSjyhSVoBMak6ypmbABWmJEmPwtKhhw?e=tbthMD

• Processed Dataset: https://huggingface.co/datasets/albertklorer/readingbank

Virtual Environment

• macOS/Linux:

  python -m venv .venv
  source .venv/bin/activate
  

• Windows (PowerShell):

  python -m venv .venv
  .venv\Scripts\activate
  

Install

pip install -r requirements.txt

Run

python main.py --dataset_directory /path/to/ReadingBank --hub_repo your-username/readingbank --hf_token hf_your_token_here