diff --git a/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py b/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py index a3118b8f73..a55ae4dffb 100644 --- a/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py +++ b/mteb/tasks/Retrieval/deu/GermanDPRRetrieval.py @@ -25,13 +25,13 @@ class GermanDPR(AbsTaskRetrieval): eval_splits=[_EVAL_SPLIT], eval_langs=["deu-Latn"], main_score="ndcg_at_10", - date=None, - domains=None, - task_subtypes=None, - license=None, - annotations_creators=None, - dialect=None, - sample_creation=None, + date=("2020-05-19", "2021-04-26"), + domains=["Written", "Non-fiction", "Web"], + task_subtypes=["Question answering"], + license="cc-by-4.0", + annotations_creators="human-annotated", + dialect=[], + sample_creation="found", bibtex_citation="""@misc{möller2021germanquad, title={GermanQuAD and GermanDPR: Improving Non-English Question Answering and Passage Retrieval}, author={Timo Möller and Julian Risch and Malte Pietsch}, diff --git a/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py b/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py index ba6a21e96e..2de3c2441b 100644 --- a/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py +++ b/mteb/tasks/Retrieval/deu/GermanQuADRetrieval.py @@ -31,7 +31,7 @@ class GermanQuADRetrieval(AbsTaskRetrieval): metadata = TaskMetadata( name="GermanQuAD-Retrieval", description="Context Retrieval for German Question Answering", - reference="https://www.kaggle.com/datasets/GermanQuAD", + reference="https://huggingface.co/datasets/deepset/germanquad", dataset={ "path": "mteb/germanquad-retrieval", "revision": "f5c87ae5a2e7a5106606314eef45255f03151bb3", @@ -42,13 +42,13 @@ class GermanQuADRetrieval(AbsTaskRetrieval): eval_splits=["test"], eval_langs=["deu-Latn"], main_score="mrr_at_5", - date=None, - domains=None, - task_subtypes=None, - license=None, - annotations_creators=None, - dialect=None, - sample_creation=None, + date=("2020-05-19", "2021-04-26"), + domains=["Written", "Non-fiction", "Web"], + task_subtypes=["Question answering"], + license="cc-by-4.0", + annotations_creators="human-annotated", + dialect=[], + sample_creation="found", bibtex_citation="""misc{möller2021germanquad, title={GermanQuAD and GermanDPR: Improving Non-English Question Answering and Passage Retrieval}, author={Timo Möller and Julian Risch and Malte Pietsch}, diff --git a/tests/test_TaskMetadata.py b/tests/test_TaskMetadata.py index 75f3095a48..3b5aa1e158 100644 --- a/tests/test_TaskMetadata.py +++ b/tests/test_TaskMetadata.py @@ -53,8 +53,6 @@ "TwitterHjerneRetrieval", "GerDaLIR", "GerDaLIRSmall", - "GermanDPR", - "GermanQuAD-Retrieval", "LegalQuAD", "AILACasedocs", "AILAStatutes",