Note: if you are looking for a tool to evaluate your simplification models, head over to EASSE instead.
This repository contains the original implementation of the evaluation methods presented in Reference-less Quality Estimation of Text Simplification Systems (1st Workshop on Automatic Text Adaption, INLG 2018). The version that was used at submission time is on branch submission.
You can get a quick glance on what the code does in demo/reference_less_evaluation.ipynb.
- Python 3.6
- Java
- Optional: QuEst (install in
resources/tools/quest/) - Optional: TERp (install in
resources/tools/terp/)
git clone git@github.com:facebookresearch/text-simplification-evaluation.git
cd text-simplification-evaluation
pip install -e .
pip install -r requirements.txt
pytest tests/
If you use this code, please cite:
L. Martin, S. Humeau, PE. Mazaré, E. De la Clergerie, A. Bordes, B. Sagot, Reference-less Quality Estimation of Text Simplification Systems
If you have any question, please contact the author: Louis Martin (louismartin@fb.com)
See the LICENSE file for more details.