This repository contains code, dataset and models for Urdu text simplification as described in paper SimplifyUR: Unsupervised Lexical Text Simplification for Urdu.
The source is available as a Jupyter notebook for a Python 3 kernel. Please see requirements.txt for details.
Pre-trained models including Word2Vec, Parts of Speech (PoS) tagger and Language Model (LM) are available for download. Download and extract them to root directory, SimplifyUR.
A parallel corpus of complex-simplified Urdu sentence-pairs is the Data folder.
If you use this tool in any of your work, please cite below paper.
SimplifyUR: Unsupervised Lexical Text Simplification for Urdu
Copyright (c) 2019 CSaLT, ITU
Code licensed under the MIT License: http://opensource.org/licenses/MIT. Data licensed under CC-BY 4.0: https://creativecommons.org/licenses/by/4.0/