Efficient-NLP-multistage-training

Source code of the IJCAI 2023 paper: "Efficient NLP Model Finetuning via Multistage Data Filtering"

Main Organization of the Code

We provide the three stage training python scripts for glue/amazon/ag news datasets. dataset.py is for data preprocessing.

Reference

If you find the code useful, please cite the following papers:

Efficient NLP Model Finetuning via Multistage Data Filtering. Xu Ouyang, Shahina Mohd Azam Ansari, Felix Xiaozhu Lin, Yangfeng Ji. the 32nd International Joint Conference on Artificial Intelligence (IJCAI 2023)

@inproceedings{ouyang2023efficient,
  title={Efficient NLP Model Finetuning via Multistage Data Filtering},
  author={Ouyang, X and Ansari, S and Lin, F and Ji, Y},
  booktitle={INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE},
  year={2023}
}