code and data for AAAI2023 paper "Adversarial Word Dilution as Text Data Augmentation in Low-Resource Regime"
A sample script to prepare a dataset:
python datasets/create_fsl_dataset.py -datadir datasets/stsa -num_train 10 -num_dev 10 -lower
run.sh: A sample bash script to run on STSA dataset, remember to change the path to root directory.