DWD

该数据集是从Wikidata中筛选出的大型子集，排除了特定的领域类别，涵盖了超过3700万个条目，并包含了大约1.66亿条陈述。与其他基准测试相比，DWD的数据量大了好几个数量级，提供了一个真实评估的数据集。其规模宏大，条目数量超过3700万，任务主要是链接预测。This dataset is a large curated subset extracted from Wikidata, with specific domain categories excluded. It includes over 37 million entries and approximately 166 million statements. Compared with other benchmark datasets, DWD has a scale several orders of magnitude larger, providing a realistic dataset for evaluation. Given its substantial size, the primary task supported by this dataset is link prediction.