HGAT

tensorflow implementation for HGAT

Requirements

tensorflow-gpu==1.13.1
numpy==1.15.4
sklearn
python 3.7

Usage

• generate bert embedding by bert-as-service
  
• generate graph matrix(sub-question legnth: 15, sub-answer length: 20)

python graph.py

• generate train and test.pkl (can modify settings in Setting.py)

python Dataset.py

• train

python train.py

• test

python train.py --train=False