metadata

library_name: PyLaia
license: mit
tags:
  - PyLaia
  - PyTorch
  - atr
  - htr
  - ocr
  - historical
  - handwritten
metrics:
  - CER
  - WER
language:
  - la
  - fr
datasets:
  - Teklia/Alcar
pipeline_tag: image-to-text

PyLaia - HOME-Alcar

This model performs Handwritten Text Recognition in Latin on medieval documents.

Model description

The model was trained using the PyLaia library on two medieval datasets:

Himanis (French);
HOME-Alcar (Latin).

Training images were resized with a fixed height of 128 pixels, keeping the original aspect ratio.

An external 6-gram character language model can be used to improve recognition. The language model is trained on the text from the HOME-Alcar training set.

Evaluation results

On HOME-Alcar text lines, the model achieves the following results:

set	Language model	CER (%)	WER (%)	lines
test	no	8.35	26.15	6,932
test	yes	7.85	23.20	6,932

How to use?

Please refer to the PyLaia documentation to use this model.

Cite us!

@inproceedings{pylaia2024,
    author = {Tarride, Solène and Schneider, Yoann and Generali-Lince, Marie and Boillet, Mélodie and Abadie, Bastien and Kermorvant, Christopher},
    title = {{Improving Automatic Text Recognition with Language Models in the PyLaia Open-Source Library}},
    booktitle = {Document Analysis and Recognition - ICDAR 2024},
    year = {2024},
    publisher = {Springer Nature Switzerland},
    address = {Cham},
    pages = {387--404},
    isbn = {978-3-031-70549-6}
}