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

# Pylia - Esposalles handwritten text recognition

This model performs Handwritten Text Recognition in Catalan (19th century).

## Model description

The model has been trained using the PyLaia library on the [Esposalles](https://rrc.cvc.uab.es/?ch=10&com=introduction/) dataset.

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

| split | N lines | 
| ----- | ------: | 
| train |  2,328  |
| val   |    742  |
| test  |    757  |

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

## Evaluation results

The model achieves the following results:

| set   | Language model | CER (%)    | WER (%) | N lines   |
|:------|:---------------|:----------:|:-------:|----------:|
| test  | no             | 0.76       | 2.62    |      757  |
| test  | yes            | 1.04       | 3.38    |      757  |

## How to use

Please refer to the [documentation](https://atr.pages.teklia.com/pylaia/).

## Cite us

```bibtex
@inproceedings{pylaia-lib,
    author = "Tarride, Solène and Schneider, Yoann and Generali, Marie and Boillet, Melodie and Abadie, Bastien and Kermorvant, Christopher",
    title = "Improving Automatic Text Recognition with Language Models in the PyLaia Open-Source Library",
    booktitle = "Submitted at ICDAR2024",
    year = "2024"
}
```