The code of ICDM'21 paper "Deep Generation of Heterogeneous Networks"
The code has been tested under Python 3.7 and Pytorch 1.7.1
The data can be accessed here.
For the real-world heterogneous graph inference:
python hgen.py
For the synthetic heterogneous graph inference:
python hgen_syn.py
All the hyperparameters are tuned and dataset have been attached. Run the following code to see available parameters that can be passed in:
python hgen.py -h
The following is the dependency of the project:
python=3.8.1
pytorch=1.10
networkx=2.5
numpy=1.20.2
In terms of generating your own latent embedding file, you could use (hin2vec_emb_gen.py) as a reference. The actual model of Hin2Vec is maintained here.