LiRo (Limba Română) provides an open benchmark for Romanian language tasks, available at lirobenchmark.github.io/.
The project keeps track of performance of different published models on the datasets and tasks listed below. This allows easy comparison of different models and monitoring progress on these tasks and datasets over time.
We encourage submissions of both new models on existing tasks as well as new datasets! The submission process is open and transparent, just open a new issue with the appropriate template and fill in your model details and performance on your target tasks.
This project is being maintained by volunteers from #AIRomania ❤️. If you use results in your work, please cite:
@inproceedings{
liro2021,
title={LiRo: Benchmark and leaderboard for Romanian language tasks},
author={Stefan Daniel Dumitrescu and Petru Rebeja and Beata Lorincz and Mihaela Gaman and Andrei Avram and Mihai Ilie and Andrei Pruteanu and Adriana Stan and Lorena Rosia and Cristina Iacobescu and Luciana Morogan and George Dima and Gabriel Marchidan and Traian Rebedea and Madalina Chitez and Dani Yogatama and Sebastian Ruder and Radu Tudor Ionescu and Razvan Pascanu and Viorica Patraucean},
booktitle={Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 1)},
year={2021},
url={https://openreview.net/forum?id=JH61CD7afTv}
}