The code for our paper "Label-Specific Feature Augmentation for Long-Tailed Multi-Label Text Classification”.
- python==3.7.10
- numpy==1.21.2
- scipy==1.7.2
- scikit-learn==0.22
- pytorch==1.10.0
- gensim==4.1.2
- nltk==3.6.5
- tqdm==4.62.3
If you want to use other datasets, please use the following command for preprocessing first:
preprocess.pyPlease confirm the corresponding configuration file. Make sure the data path parameters (train_texts, train_labels and etc.) are right in:
main.pyTrain and evaluate as follows:
python main.py