README.md

STS and STR Evaluation Datasets

To load datasets for evaluation with script get_data.py, some datasets need to be downloaded first. In the data/sts_str folder, you can run the following commands:

STS17 data for Czech, German and French

git clone https://gitlab.com/tigi.cz/cross-lingual-sts.git

The datasets should now be in data/sts_str/cross-lingual-sts.

Karde-NLU

git clone https://github.com/lksenel/Kardes-NLU.git

The datasets should now be in data/sts_str/Kardes-NLU.

STR24

git clone https://github.com/semantic-textual-relatedness/Semantic_Relatedness_SemEval2024.git

The datasets should now be in data/sts_str/Semantic_Relatedness_SemEval2024.