torchtext使用 ValueError: invalid literal for int() with base 10: ''wen'ti

batch进行迭代时候报这个错误，试了好久都不知道是哪里出错了

C:\ToolsAnaconda3\lib\site-packages\torchtext\data\field.py in <listcomp>(.0)
    353             if not self.sequential:
    354                 arr = [numericalization_func(x) if isinstance(x, six.string_types)
--> 355                        else x for x in arr]
    356             if self.postprocessing is not None:
    357                 arr = self.postprocessing(arr, None)

ValueError: invalid literal for int() with base 10: ''

我也有这个问题，请问题主解决了嘛