Due to the size limit of GitHub, the code and all the datasets used can be found at (https://osf.io/9w7ks/?view_only=c96e5611809845568b8c0d81acdb2b46)
-
Python 3.6
-
Tensorflow r1.0.1
-
CUDA 7.5+
All the defult parameters are listed in 'train.py' inside the SeqGAN folder. To train SeqGAN, run
' $ python train.py '
The generated data should appear inside the 'save' folder, which can be used to further fine-tune HateBERT.