Printed Arabic Base Model Trained on the OpenITI Corpus

This is a text recognition model trained on the OpenITI dataset of printed Arabic-script text available at [0] in its state of 2022-09-03. It encompasses real world Arabic (~23k lines) material in a variety of typefaces augmented by synthetic data in the Tahoma (600 lines) typeface. The model has been obtained by fine-tuning the Arabic-script base model [1] on the purely Arabic-language subset of the corpus. As the model is trained on a variety of highly diverse typefaces it is mostly intended as a base model for fine-tuning more specific models from it. In line with this it has not been extensively verified or optimized. The ground truth was lightly normalized to NFD but is otherwise untouched. [0]: https://github.com/OpenITI/arabic_print_data.git [1]: 10.5281/zenodo.7050270