Hello Wikidata folks,
I would like to bring your attention to an open source dataset I've been
developing called the Kensho Derived Wikimedia Dataset (KDWD). It's a
cleaned English subset of Wikipedia/Wikidata with 2.3B tokens, 5.3M pages,
51M nodes, and 120M edges. More details are available here
https://blog.kensho.com/announcing-the-kensho-derived-wikimedia-dataset-5d1…
best,
-Gabriel