AttributeError: 'str' object has no attribute 'tolist'

[mading0817]

Hi @jakeywu,

Thanks for your great project, the code is quite readable and efficient!
I could successfully run the train_model.py and the model can be saved successfully as well.
The problem is when I tried to run predict.py, it comes out a following error message:
Traceback (most recent call last): File "E:/chinese_ner-master/chinese_ner-master/predict.py", line 70, in <module> predict = ner.batch_predict_ner() File "E:/chinese_ner-master/chinese_ner-master/predict.py", line 55, in batch_predict_ner sentence_ner.append(self.source_data[i].tolist()[j]) AttributeError: 'str' object has no attribute 'tolist'

Thanks again!

[mading0817]

by the way, I was using tensorflow-gpu 1.10.0 with python 3.6.5

[mading0817]

Hi,

I had to change line 59 in predict.py from:
sentence_ner.append(self.source_data[i].tolist()[j])
to
sentence_ner.append(self.source_data[i][j])

Then it could run successfully, but the result might be wrong:
[ [ "美", "国", "团", "购" ], [ "美", "国", "团", "购" ] ]

decode_tags: decode: [[2 3 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
[2 3 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]``

The demo only shows how it recognize ORG, could you show me about how do I change for multiple entities recognition?

Thanks!!