emnlp2017-bilstm-cnn-crf
emnlp2017-bilstm-cnn-crf copied to clipboard
Adding new word features (prefix/suffix) for training
As shown this is an improvement done for a same like architecture as discussed in your paper.
If I need to include those additional word features how should I change the code in BiLstm.py? should I include for each prefix and suffix feature, with a input layer and an embedding layer? I will be grateful for your suggestions