SloBench

Slovenian Natural Language Processing Benchmark

SloBENCH is an evaluation platform for benchmarking the development of Slovene natural language processing technologies. The main goal of SloBENCH evaluation is that evaluation data labels are not publicly published. Additionally, submission limitation are defined to avoid overfitting of models. Still we ask all the submitters to respect ethics and to not manually annotate the test data.

The system was developed as one of Clarin.si's 2021 projects and was proposed by Slavko Žitnik, Marko Robnik-Šikonja and Simon Krek. The implementation was done by Frenk Dragar.