We often deal with label errors in datasets, but no common framework exists to support machine learning research and benchmarking with label noise. Announcin...