The framework of this project is constructed according to pytorch-worker, you can visit the project for more details about usage.

In the directory config/element, there exists three different directories labor,divorce,loan representing three parts of the dataset. Besides, in each directory, there exists several config files representing different models:

• bidaf.config: BIDAF.
• chinese.config: BERT presented by Google.
• cnn.config: TextCNN.
• dpcnn.config: DPCNN.
• lstm.config: LSTM.
• ms.config: BERT pretrained on civil documents.