LR-CNN

Source codes for paper: CNN-Based Chinese NER with Lexicon Rethinking

Requirement:

Python 3.6 Pytorch 0.4.1

Input format:

CoNLL format, with each character and its label splited by a whitespace in a line. The "BMES" tag scheme is prefered.

别 O  
错 O  
过 O  
邻 O  
近 O  
大 B-LOC  
鹏 M-LOC  
湾 E-LOC  
的 O  
湿 O  
地 O  

Pretrain embedding:

The pretrained embeddings(word embedding, char embedding) are the same with Lattice LSTM(https://github.com/jiesutd/LatticeLSTM)

Run the code:

1. Download the character embeddings and word embeddings and put them in the data folder.
2. To train/test the demo: sh train.sh / sh test.sh
3. To train/test your own data: modify the 'train.sh' or 'test.sh' file with your file path, and run the shell file.