大神你好，请教一些关于报错的问题

[ljx761795750]

大神你好，我的配置都没问题，数据是韩文中文和英文，数据都是用B/I/E-ORG标记的，在train期间是正常的，但是decode就出错，我的raw_dems数据是三种文字，后边标记全是O，请问下面报错是什么原因啊。我的python3.8和pytorch1.6.。。。
下面是我decode后报错的问题，请问 这是版本的问题么？还是其他的问题。谢谢。

`DATA SUMMARY END.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Traceback (most recent call last):
File "/tmp/test_ljx/2/NCRFpp-master/utils/alphabet.py", line 49, in get_index
return self.instance2index[instance]
KeyError: 'O'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "main.py", line 562, in
data.generate_instance('raw')
File "/tmp/test_ljx/2/NCRFpp-master/utils/data.py", line 306, in generate_instance
self.raw_texts, self.raw_Ids = read_instance(self.raw_dir, self.word_alphabet, self.char_alphabet, self.feature_alphabets, self.label_alphabet, self.number_normalized, self.MAX_SENTENCE_LENGTH, self.sentence_classification, self.split_token)
File "/tmp/test_ljx/2/NCRFpp-master/utils/functions.py", line 109, in read_instance
label_Ids.append(label_alphabet.get_index(label))
File "/tmp/test_ljx/2/NCRFpp-master/utils/alphabet.py", line 56, in get_index
return self.instance2index[self.UNKNOWN]
KeyError: ''
`

[jiesutd]

可能需要检查一下数据格式