One participating team was so kind as to provide additional test data for the various categories.
We have looked into their samples and believe the quality is pretty good. If you are in need of further test data, you may use the samples provided. Using those samples alone will not result in a better or worse evaluation score. Please be aware that the jury might have classified some results in the sample data differently, but in most cases agree with the categorization.
If you like these examples and think it only would be fair if you provided some by yourself, feel free to open a pull request for this repository.
