Konstantin Rusch comments

Results 6 comments of


                                            Konstantin Rusch

Reproducibility

We presented the one best result in the paper. I only tested it on one specific GPU with one specific pytorch version (note that seeds change throughout different machines and...

Code Question

Thanks for reaching out. It is a multi-layer GNN, however, in our case we share the same parameters among the different layers. That's why we do the for-loop over the...

No, it's absolutely not! You can extend it to using different parameters among different layers by simply: ``` self.convs = nn.ModuleList() for i in range(nlayers): self.convs.append(GCNConv(nhid, nhid)) ``` and then...

The results on ZINC dataset is much worse than that in the paper

I re-ran my code on a different machine and could reproduce the result we stated in the paper. Can you elaborate more on the issue? What MAE do you obtain...

ARM64 torch_cluster.radius bug

I'm having the same issue. Any updates on that?

CUDA extension: unable to compile

My apologies for the delayed reply. Were you able to compile it after all?