Is all datasets in /data are supported ?

[Zhaohan]

Hi Dear Author,

This looks a very good implementation in using GCN for graph-level classification (may also called inductive). Besides the sample running you shown in jupyter notebook, are the other datasets are also supported in running? If yes, could I ask which part should modify to change the calling of datasets? Thanks!

[Boris Knyazev]

Hi,
The DataReader class can in principle read any dataset in the TU format (see datasets here https://ls11-www.cs.tu-dortmund.de/staff/morris/graphkerneldatasets).
You need to download the dataset, put it in the ./data/ folder and change line dataset = 'proteins'.
But I think it's better to use the python script in my repo (https://github.com/bknyaz/graph_nn/blob/master/graph_unet.py), because I tested it more and for different datasets. It has the --dataset argument that you can use to specify any TU dataset.

[Zhaohan]

Got, thanks for response