Skip to content

Dataset · fhswf

German handwriting samples for OCR

fhswf/german_handwriting

About 10k images of handwritten German text with transcripts for training recognition systems.

What's inside

The dataset contains around 10,000 pairs of handwritten German text images and corresponding text labels. Samples were collected from 15 different authors based on school and university notes. The dataset was created as part of a handwriting recognition project at FH-SWF University.

What you can do with it

  • Training or fine-tuning OCR models (image-to-text) to recognize German handwriting
  • Testing multimodal models on the task of digitizing handwritten notes
  • Evaluating the robustness of text recognition systems to different handwriting styles
Tasks
Image → text
Languages
German
Size
10K–100K rows
Downloads this month
482
Likes
19

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("fhswf/german_handwriting")