This repository implements an LSTM-CRF model for named entity recognition. The model is same as the one by Lample et al., (2016) except we do not have the last tanh
layer after the BiLSTM.
- DyNet 2.0
- Python 3
For DyNet, CPU is sufficient for the speed.
We conducted experiments on the CoNLL-2003 dataset.
Dataset | Precision | Recall | F1 score |
---|---|---|---|
CoNLL-2003 | 90.57 | 91.26 | 90.91 |
- Put the Glove embedding file (
glove.6B.100d.txt
) underdata
directory - Simply run the following command and you can obtain results comparable to the benchmark above.
python3 main.py
- Use Elmo and Bert as embeddings. For now, we use Glove.
- Online demo for testing
- Train the model on larger corpus (i.e., OntoNotes 5.0).
Lample, Guillaume, Miguel Ballesteros, Sandeep Subramanian, Kazuya Kawakami, and Chris Dyer. "Neural Architectures for Named Entity Recognition." In Proceedings of NAACL-HLT, pp. 260-270. 2016.