Hugging Face's logo

Datasets:
k8mpass-fraktur-series
/
fraktur-baltic-corpus

Tasks:
Translation
Modalities:
Text
Formats:
json
Sub-tasks:
language-modeling
Languages:
German
Russian
English
Size:
< 1K
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
original_fraktur
stringclasses
3 values
normalized_german
stringclasses
3 values
translation_en
stringclasses
3 values
translation_ru
stringclasses
3 values
translation_et
stringclasses
3 values
translation_sv
stringclasses
3 values
translation_fi
stringclasses
3 values
translation_da
stringclasses
3 values
Der Qerlafter
Der Verfasser
The author
Автор
Autor
Författaren
Kirjoittaja
Forfatteren
Das ift eine Probe
Das ist eine Probe
This is a sample
Это пример
See on näidis
Detta är ett exempel
Tämä on esimerkki
Dette er et eksempel
Narva war eine wichtige Festung
Narva war eine wichtige Festung
Narva was an important fortress
Нарва была важной крепостью
Narva oli tähtis kindlus
Narva var en viktig fästning
Narva oli tärkeä linnoitus
Narva var en vigtig fæstning

Fraktur Baltic Corpus

Fraktur Baltic Corpus is a multilingual dataset based on historical German-language books printed in the Russian Empire during the 18th–19th centuries, primarily in Fraktur typeface.

Each entry in the dataset contains:

  • Raw OCR text from historical Fraktur sources
  • Normalized German version
  • Translations into seven languages: English, Russian, Estonian, Swedish, Finnish, Danish, and Modern German

Volume 1: Hansen, Geschichte der Stadt Narva (1858)

This initial release includes the "Einleitung" (introduction) section of the 1858 book Geschichte der Stadt Narva by H. J. Hansen, a Danish-German ethnographer. The dataset includes:

  • 87 aligned records
  • OCR text (Fraktur)
  • Normalized German
  • 7 parallel translations

Format

Each .jsonl and .csv entry includes the following fields:

  • original_fraktur
  • normalized_german
  • translation_en
  • translation_ru
  • translation_et
  • translation_sv
  • translation_fi
  • translation_da

License

CC BY-SA 4.0
You are free to use, adapt, and redistribute with proper attribution.

Author & Contact

This dataset is created by k8mpass, an independent open-source initiative based in Estonia.
Feel free to cite, fork, or contact for collaborations.

📥 Download instructions

  • Use dataset.csv to explore or visualize the corpus.
  • To use the full dataset for machine learning training:
    1. Download data.zip
    2. Unzip it to extract the file data.jsonl
    3. Load data.jsonl using your preferred JSONL reader or training pipeline
  • File is encoded in UTF-8 and follows JSON Lines format (one JSON object per line).
Downloads last month
9
Size of downloaded dataset files:
1.03 kB
Size of the auto-converted Parquet files:
5.35 kB
Number of rows:
3