Context This dataset contains a little more than 8 Million sentences from the German Wikipedia. Data was obtained via the [LinguaTools Monolingual Dumps](https://linguatools.org/tools/corpora/wikipedia-monolingual-corpora/), filtered to only keep articles with more than 35 links pointing to them and stripped of XML (and other junk). Articles were then split into sentences using [NNSplit](https://github.com/bminixhofer/nnsplit). Acknowledgements Obviously: I did almost no work here, all credit goes to [Wikipedia](https://www.wikipedia.org/) and [Linguatools](https://linguatools.org/). Inspiration Some time ago I didn't find any easy way to access the (cleaned) data from German Wikipedia for a project with self-supervised training. I just revisited it and thought I'd put the data up here, I'm not sure if there is still a need for it but it can't hurt I guess :)