Source code of our paper presents in EMNLP 2019. https://www.aclweb.org/anthology/D19-1345/
A list of parameters required to run the code can be obtained in the following way.
python3 train.py -h
Examples:
python3 train.py --dataset r8 --ngram 3
python3 train.py --dataset r52 --ngram 3
python3 train.py --dataset oh --ngram 4
Note: You need to put a pre-trained Glove model file named glove.6B.200d.vec.txt in the root directory.
You get the file from https://nlp.stanford.edu/projects/glove/.
You need to add a line at the top of the file with the following value:
400000 200