NLP_textcnn

(10条消息) 文本分类系列(1):textcnn及其pytorch实现_u014514939的博客-CSDN博客

   问题1：UnicodeDecodeError: 'gbk' codec can't decode byte 0x80 in position 111: illegal multibyte sequence

解决方法：python打开文本，改变文件的编码方式在

file_object = open('data/stopwords.txt',encoding='utf-8')添加encoding='utf-8'

问题2：

AttributeError: module 'torchtext.data' has no attribute 'Field'

解决方法：将 from torchtext import data 改为from torchtext.legacy import 

问题3：RuntimeError: set_storage_offset is not allowed on a Tensor created from .data or .detach().

解决方法：将

feature.data.t_(), target.data.sub_(1)改为

feature = feature.data.t()
target = target.data.sub(1)

Ask Question