We find that, although our aligned dataset is much harder than previous datasets, transformer-based models recognize the humorous sentence ...
確定! 回上一頁