GraphGAN

• This is the implementation of AAAI2018. GraphGAN

code structure

• data : includes the graph file

• pre_train : includes the pre_train embeddings file that generated by other methods, such as deepwalk, SDNE, LINE, since gan is difficult to be trained.

• results : includes the evaluation results, you can see the results improving with the training progressing.

• src : includes the source codes

  ​

requirements:

• tensorflow
• tqdm(display the running progress)

data

undirected graph whose file format is as follows(from 0 to N(number of nodes)):

0 1

3 2

...

Run

cd src/GraphGAN

python graph_gan.py