/CRE-wih-relation-information

Primary LanguagePython

CRL

Implementation of the research paper Consistent Representation Learning for Continual Relation Extraction (Findings of ACL 2022)

Framwork of consistent representation learning

Memory Replay Architecture

Dependencies

Use anaconda to create python environment:

conda create --name yourname python=3.8
conda activate yourname

Install Pytorch (suggestions>=1.7) and related environmental dependencies:

pip install -r requirements.txt

Pre-trained BERT weights:

  • Download bert-base-uncased into the datasets/ directory [google drive].

Run the Code

python run_continual.py --dataname FewRel

Citation

Please cite our paper if you find our work useful for your research:

@misc{zhao2022consistent,
      title={Consistent Representation Learning for Continual Relation Extraction}, 
      author={Kang Zhao and Hua Xu and Jiangong Yang and Kai Gao},
      year={2022},
      eprint={2203.02721},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}