README.md

Reference

If you use the code, please cite the following paper:

@article{DBLP:journals/nca/QinWZLD23,
  author       = {Shaowei Qin and
                  Hao Wu and
                  Lihua Zhou and
                  Jiahui Li and
                  Guowang Du},
  title        = {Learning metric space with distillation for large-scale multi-label
                  text classification},
  journal      = {Neural Comput. Appl.},
  volume       = {35},
  number       = {15},
  pages        = {11445--11458},
  year         = {2023}
}

Requirements

• python==3.8.8
• tensorflow==2.7.0

Datasets

• Reuter
• EUR-Lex
• Wiki10-31K
• CiteULike-t
• Download the GloVe embedding (840B,300d) (https://nlp.stanford.edu/projects/glove/)

Train and Test

Run main.py for train and test datasets with tokenized texts as follows:

python main.py --data 'Wiki10-31K' --la 1.0 --ba 1.0 --emb_num 256 --n_negative 1000