metadata
configs:
- config_name: default
data_files:
- split: corpus
path: corpus/data-00000-of-00001.arrow
- split: queries
path: queries/data-00000-of-00001.arrow
license: cc-by-4.0
language:
- de
source_datasets:
- deepset/germanquad
This dataset is derived from the GermanQuAD dataset.
This dataset takes the testset and represents it as a corpus in the BEIR information retrieval benchmark format. The corresponding qrels can be found here.
Full credit for the original dataset goes to the authors of the GermanQuAD dataset.