nairouz/ft-vgae

IJCAI 2022- Escaping Feature Twist: A Variational Graph Auto-Encoder for Node Clustering

ft-vgae's Issues

Unable to run

Hi, I'm unable to run this, as there are no pretrained weights that the model needs and they are not available at the mentioned https://github.com/zfjsail/gae-pytorch/ repo either, and their training code does not save a model either.

Could you please provide either the training weights directly (preferable) or a code that trains the model and saves weights usable by this code?

I did try running their code and saving a model, but it expects different layers:

RuntimeError: Error(s) in loading state_dict for FT_VGAE:
        Missing key(s) in state_dict: "base_gcn.weight", "gcn_mean.weight", "gcn_logsigma2.weight", "assignment_1.cluster_centers".
        Unexpected key(s) in state_dict: "gc1.weight", "gc2.weight", "gc3.weight".

So is that repo not being used anymore?

Add requirements.txt

For reproducibility, it would be useful to mention the version of Python used and to include a requirements.txt file in the repository.