Datasets:

wkrl
/

cord

Dataset Viewer

View in Dataset Viewer

Viewer

The dataset viewer is not available for this dataset.

The dataset viewer doesn't support this dataset because it runs arbitrary python code. Please open a discussion in the discussion tab if you think this is an error and tag @lhoestq and @severo.

Error code:   DatasetWithScriptNotSupportedError

Need help to make the dataset viewer work? Open a discussion for direct support.

Dataset Card for CORD (Consolidated Receipt Dataset)

Dataset Summary

[More Information Needed]

Supported Tasks and Leaderboards

[More Information Needed]

Dataset Structure

Data Instances

[More Information Needed]

Data Fields

{
  "id": datasets.Value("string"),
  "words": datasets.Sequence(datasets.Value("string")),
  "bboxes": datasets.Sequence(datasets.Sequence(datasets.Value("int64"))),
  "labels": datasets.Sequence(datasets.features.ClassLabel(names=_LABELS)),
  "images": datasets.features.Image(),
}

Data Splits

train (800 rows)
validation (100 rows)
test (100 rows)

Dataset Creation

Licensing Information

Creative Commons Attribution 4.0 International License

Citation Information

@article{park2019cord,
  title={CORD: A Consolidated Receipt Dataset for Post-OCR Parsing},
  author={Park, Seunghyun and Shin, Seung and Lee, Bado and Lee, Junyeop and Surh, Jaeheung and Seo, Minjoon and Lee, Hwalsuk}
  booktitle={Document Intelligence Workshop at Neural Information Processing Systems}
  year={2019}
}

Contributions

Thanks to @clovaai for adding this dataset.