Lukaszl/pl-newspaper-pages-ocr-dataset-100

Polish Press Pages OCR Dataset 100

A small public image-only OCR dataset containing 100 Polish-language press / magazine-style page images sampled from a multilingual document collection.

This dataset is designed as a lightweight evaluation sample for:

• OCR models
• VLM-based document understanding
• testing OCR robustness on Polish multi-column and press-style layouts

Contents

• 100 JPG images
• Polish-language page images
• press / magazine-style layouts, including:
  • article pages
  • multi-column text layouts
  • pages with embedded images
  • pages with side blocks and highlighted sections
  • advertisement-style pages
  • editorial / magazine-like compositions

What is NOT included

• ground-truth transcriptions
• OCR outputs
• benchmark scores
• dataset splits

Purpose

This is not a full benchmark.

It is a compact dataset sample intended for:

• testing OCR robustness on Polish press-style pages
• quick OCR sanity checks
• evaluating performance on non-trivial reading-order layouts
• checking how models handle mixed typography and page composition

Source

This dataset is a derived Polish-only subset of the original upstream dataset:

• AlekseyScorpi/docs_on_several_languages https://huggingface.co/datasets/AlekseyScorpi/docs_on_several_languages

The upstream dataset is multilingual. This repository contains a Polish-only image subset selected for OCR evaluation purposes.

License / attribution

This repository is a derived public subset of the upstream source.

Please refer to the original dataset page for licensing and attribution terms:

• https://huggingface.co/datasets/AlekseyScorpi/docs_on_several_languages

If you use this subset, please also attribute the original upstream dataset where appropriate.

Maintainer

clearOCR https://clearocr.com

Notes

• image-only dataset
• Polish-language subset
• press / magazine-style pages
• intended for evaluation, not training
• useful as a compact OCR stress-test sample