CFER

Paper: Coarse-to-Fine Entity Representations for Document-level Relation Extraction (URL)

Requirements

• Pytorch (1.6.0)
• numpy (1.19.4)
• tqdm (4.50.2)
• transformers (4.6.1)
• spacy (2.3.2)

Downloading Data

You can download the required data from here. After you download data.zip, unzip it and put it to the root directory of this project. Besides the datasets, we also provide some preprocessing results (see data/adj/ and data/path) for saving time.

Specifying Configuration

Before training, you can edit code/config.py to specify the configurations, including filepath information, and hyper-parameters. If you want to reproduce our results reported in the paper, you can use the reported hyper-parameters, and keep other hyper-parameters unchanged.

Training and Evaluating

1. Change the working directory to the root directory of this project.
2. Run python3 code/main.py.

Citation

If you use this code for your research, please kindly cite our paper:

@article{dai2020cfer,
        author    = {Damai Dai and
                    Jing Ren and
                    Shuang Zeng and
                    Baobao Chang and
                    Zhifang Sui},
        title     = {Coarse-to-Fine Entity Representations for Document-level Relation Extraction},
        journal   = {CoRR},
        volume    = {abs/2012.02507},
        year      = {2020},
        url       = {https://arxiv.org/abs/2012.02507}
}

Contact

This project is supported by Jing Ren. If you have any problems, please contact us via the following e-mail addresses.

Jing Ren: rjj@pku.edu.cn

Damai Dai: daidamai@pku.edu.cn