:sparkles: EPiDA :sparkles:

Official Code for 'EPiDA: An Easy Plug-in Data Augmentation Framework for High Performance Text Classification'

NAACL 2022 Accepted Paper

:e-mail: Contact :e-mail:

Feel free to contact me if you have any problems! [email protected]

:fire: 2. Quick Start :fire:

Suppose you have already get a dataset, you can modify the code in train_sst_epida_eda.py or train_irony_epida_eda.py for quick start.

python3 train_irony_epida_eda.py

:beers: Data Access :beers:

The SST dataset can be achieved via SUB2.

The corpus used in major paper can be downloaded by checking the links given in the paper or email.

:satisfied: Citation :satisfied:

If you find this project is useful for your research, please cite:

@article{zhao2022epida,
  title={EPiDA: An Easy Plug-in Data Augmentation Framework for High Performance Text Classification},
  author={Zhao, Minyi and Zhang, Lu and Xu, Yi and Ding, Jiandong and Guan, Jihong and Zhou, Shuigeng},
  journal={arXiv preprint arXiv:2204.11205},
  year={2022}
}