Repository for holding source code of paper Generating Adversarial Examples for Holding Robustness of Source Code Processing Models.
MHM
├── README.md
├── mhm
├── LSTMClassifier # training LSTM model on poj104 classification task.
├── mhm-astnn # adversarial attack on LSTM classifier.
├── mhm-lstm # training ASTNN model and adversarial attack on it.
└── poj104 # filelist of poj104 dataset.
python 3.5 with package requirements:
jupyterlab
numpy
pytorch
pandas
pycparser
gensim
tensorflow
Transformer Model is built based on The Annotated Transformer.
ASTNN Model is built based on ASTNN.