this is implementation for lightgcn model. In this version, i use dummy data and not use batch training.

for large scale dataset, switch scale branch

python light_gcn.py