mul-BERT, the official score on the SemEval 2010 Task 8 dataset is up to 90.72 (Macro-F1).
Primary LanguagePython