The HIPE-eval organisation hosts code, datasets and evaluation tools for the HIPE series of shared tasks on historical document processing. Originally centred on named entity processing with HIPE-2020 and HIPE-2022, HIPE-eval is now expanding to address additional challenges found in noisy, heterogeneous, and multilingual historical documents, including relation extraction and OCR post-correction.

Although HIPE first stood for 'Identifying Historical People, Places and other Entities', the initiative has grown into a broader umbrella for shared tasks in historical document processing - but we keep the HIPE name 😊.

🚀 HIPE-2026: Person–Place Relation Extraction, Evaluation Lab hosted at CLEF 2026.

• 👉 Registration is open until April 23, 2026!
• HIPE-2026-data – data of the HIPE 2026 shared task.
• HIPE-2026-eval – evaluation scripts, baselines, submissions and results.

🚀 HIPE-OCRepair: LLM-assisted OCR post-correction, Competition hosted at ICDAR-2026

• 👉 Registration is open until March 23, 2026!
• HIPE-OCRepair-2026-data: for public data releases both on Github here and on Hugging Face huggingface/hipe/hipe-ocrepair-data
• HIPE-OCRepair-2026-baseline.
• HIPE-OCRepair-2026-eval: for scorer, leaderboard app, and ultimately system submissions for reproduction; both on Github and huggingface/hipe/hipe-ocrepair-2026-eval
• HIPE-OCRepair-2026-submissions: huggingface/hipe/hipe-ocrepair-2026-submissions. To receive system submission in private repositories from the HF leaderboard.

🏁 ✔️ HIPE-2022: Named Entity Recognition and Linking, Evaluation Lab hosted at CLEF-2022.

• HIPE-2022-data – data of the HIPE 2022 shared task.
• HIPE-2022-eval – all you need to reproduce the HIPE-2022 evaluation campaign results.
• HIPE-2022-baseline – baseline models for NERC for HIPE-2022.

🏁 ✔️ HIPE-2020: Named Entity Recognition and Linking, Evaluation Lab hosted at CLEF-2020.

• CLEF-HIPE-2020-eval – all you need to reproduce the HIPE-2020 evaluation campaign results (in Impresso GH organisation).

📚 Other tools and resources

• HIPE-scorer – Python module for evaluating NERC and NEL systems (CLI-based; Hugging Face integration planned).
• HIPE-pycommons – Utility functions to manipulate HIPE-formatted data and compute basic statistics.
• For more information on shared tasks results, check the publications in the 'References' section of each shared task website.
• Visit the HIPE Zenodo organisation.

How can I contribute?
Feel free to use the scorer, datasets and evaluation material (please respect the licences and cite the relevant papers).
If you find errors or have improvements to propose, issues and pull requests are welcome!

What's to come?
We hope to be able to offer another edition of HIPE, and to publish useful data again, so stay tuned!

Funding
With the exception of HIPE-2022, HIPE evaluation campaigns are carried out in the context of the Impresso - Media Monitoring of the Past projects. The first project (2017-2021) was funded by the Swiss National Science Foundation under grant No. CRSII5_173719 and the second project (2023-2027) by the SNSF under grant No. CRSII5_213585 and by the Luxembourg National Research Fund under grant No. 17498891.

Organisation
To date, HIPE evaluation campaigns have been organised jointly by the EPFL Digital Humanities Laboratory and the UZH Institute of Computational Linguistics. If you would like to organise a shared task under the HIPE umbrella, feel free to get in touch!