This is the HiNER model repository. We provide detailed data set processing, model training and prediction codes. We hope you will like our work:
Due to data sharing limitations, we provide links to the datasets:
Download the data set and conduct experiments according to the relevant settings of the data loader.
BERT is used for character embedding, Chinese dictionary TX is used for lexical embedding, and NLP tools are used for syntactic embedding. For each part of the model, view the model.py file. Main.py is used for training and evaluation. The file location is replaced with your own file location.
Our code is based on bert4torch, and thanks to the work of W2NER.