I love the open-source community and have released all the codes in my publications, as well as some useful tools.

  • NCRF++: A PyTorch implementation neural sequence labeling framework with flexiable combination of word, character, and handcrafted features with CRF layer (e.g. Bi-LSTM-CRF with char CNN). >1400 Stars.
  • YEDDA: An efficient graphical toolkit for chunk/entity/event annotation. Supports shortcuts annotation, command line annotation, system recommendation and result analysis. ACL 2018 best demo paper nomination. >500 Stars.
  • LatticeLSTM:A state-of-the-art Chinese NER system with Lattice LSTM structure. This model give the best performance on many Chinese NER datasets, gives 93.18\% (F1) on MSRA NER shared task. >900 Stars.
  • RichWordSegmentor: A state-of-the-art neural transition based Chinese word segmentor with neural rich pretraining character representation (multi-task structure). >100 Stars.
  • More code can be found in my github.