---
library_name: PyLaia
license: mit
tags:
- PyLaia
- PyTorch
- atr
- htr
- ocr
- historical
- handwritten
metrics:
- CER
- WER
language:
- fr
- la
- it
- oc
- es
datasets:
- CATMuS/medieval
pipeline_tag: image-to-text
---

# PyLaia - CATMuS/medieval

This model performs Handwritten Text Recognition in Latin/Romance on historical documents.

## Model description

The model was trained using the PyLaia library on the [CATMuS/medieval](https://huggingface.co/datasets/CATMuS/medieval) dataset.

Training images were resized with a fixed height of {dimension} pixels, keeping the original aspect ratio. Vertical lines are discarded.

| set | lines | 
| :----- | ------: | 
| train | 15,2816 |
| val   |  19,402 |
| test  |  22,590 |

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

## Training Plot


![image/png](https://cdn-uploads.huggingface.co/production/uploads/6616d277007cf0aeb2600b77/6bJ-xbwuLHc2cU3m38YZf.png)

## How to use?

Please refer to the [PyLaia documentation](https://atr.pages.teklia.com/pylaia/usage/prediction/) to use this model.

## Demo

https://huggingface.co/spaces/johnlockejrr/yolov11_pylaia_catmus