Dataset Access for ReQue

[parthpadhiar]

Currently, I can only access the Antique and ClueWeb12 datasets. Unfortunately, I have encountered challenges accessing some of the other datasets that are essential for my project, specifically Robust04, Gov2, ClueWeb09-B, and Dbpedia. I was wondering if you could guide me on how to access these datasets or if there are any alternative suggestions or sources where I might find similar datasets suitable for ReQue.
Thank you very much for your time and assistance. I’m looking forward to any guidance you may be able to provide.

[hosseinfani]

Hi @parthpadhiar
Please have a look at https://ir-datasets.com/
They collect the topics (query sets), the relevance judgments (qrels), and the actual document corpus for few.

Please note that you may not need the actual document corpus but the index that might be already created by some labs.

Though most of such indexed are build for sparse retrievers like bm25. In some exceptional cases, there is also an index for dense retrievers like colbert like for msmarco.

Nonethless, if you need the actual corpus, in https://ir-datasets.com/ explains how to obtain them. You have to purchase some of them or sign a few documents and they deliver it to you like gov2 or clueweb.