README.md

---
license: cc-by-4.0
language:
- fi
task_categories:
- image-to-text
pretty_name: Theseus Finnish OCR Tiny
dataset_info:
  features:
  - name: pdf_file
    dtype: string
  - name: page
    dtype: int32
  - name: text
    dtype: string
  - name: image
    dtype: image
  splits:
  - name: train
    num_examples: 16618
tags:
- parquet
size_categories:
- 10K<n<100K
---

# Theseus Finnish OCR Dataset

Paragraph-level OCR dataset harvested from [Theseus.fi](https://www.theseus.fi),
the Finnish repository of university of applied sciences theses.

Each record is one paragraph crop extracted from a thesis PDF, paired with the
text extracted by `pdfplumber`.

## Image Resolution

Paragraph crops are rendered at **300 DPI** (dots per inch) with 2 px
padding on each side. At 300 DPI a standard A4 page is
2481 × 3507 pixels, giving high enough resolution for training
OCR and document-understanding models.

## Fields

| Field | Type | Description |
|-------|------|-------------|
| `pdf_file` | string | Source PDF filename |
| `page` | int | Page number (1-indexed) |
| `image` | Image | Paragraph crop at 300 DPI with 2 px padding on each side |
| `text` | string | Extracted paragraph text from `pdfplumber` |

## Source

Records were harvested via the Theseus OAI-PMH endpoint and 10 PDFs were
downloaded from the DSpace bitstream API.