The first two rows are baselines from the XNLI paper and the last three rows are our results with BERT. Translate Train means that the MultiNLI training set was ...
確定! 回上一頁