Skip to content
Neural Reranking for Named Entity Recognition, accepted as regular paper at RANLP 2017
Python
Branch: master
Clone or download
Latest commit 9bb7913 Jul 15, 2017
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data init version Jul 15, 2017
model init version Jul 15, 2017
results init version Jul 15, 2017
utils init version Jul 15, 2017
A.BI.True.Full model.png init version Jul 15, 2017
readme.md init version Jul 15, 2017
reranker.py init version Jul 15, 2017

readme.md

How to run the code:

python reranker.py train DEBUG B

where
train is training and save model first, if using other word, then just load models without training.
DEBUG use small dataset. If using other word or None, use standard data.
B means LSTM with CNN model.

Data format

Generate the nbest results using neural BI-LSTM-CRF at NbestNER.
Generate the nbest results using discrete CRF model at CRF++
Data format follow the sample data

Folder

data: where the data saved model: model file saved results: training model saved utils: load data, metric file saved

Cite:

@article{yang2017neural,  
 title={Neural Reranking for Named Entity Recognition},  
 author={Yang, Jie and Zhang, Yue and Dong, Fei},  
 booktitle = {Proceedings of RANLP},
 year={2017}  
}  

Version

2017-April-4: init version

You can’t perform that action at this time.