Thomas Kipf

Yes, the adjacency matrix is "symmetric" in shape, but it is not necessarily a symmetric matrix (i.e., `A^T = A`). Maybe try a different python/numpy version or add an explicit...

This should work fine! If your adjacency matrix is already normalized (e.g. all rows sum to 1), then you might just want to skip the re-normalization (in normalize_adj) altogether.

If you have multiple edge types, you can use: https://github.com/tkipf/relational-gcn On Thu, Apr 25, 2019 at 4:14 PM Mónica Aguilar Herrero < [email protected]> wrote: > I would like to use...

Thanks for your question. I am not sure what you mean by 'online predict'- would you mind elaborating? On Fri, Feb 1, 2019 at 11:28 AM Wenkai Liu wrote: >...

Thanks for this bug report! I do not have an immediate solution for this other than suggesting to try a different Keras version (this package was developed with Keras 1.0.9...

Thanks for the pull request. I would like to keep the default split sizes as they are, as this more closely matches the dataset splits from our paper, but I...

I think this is what you’re looking for: https://keras.io/getting-started/faq/#how-can-i-obtain-the-output-of-an-intermediate-layer > On 7 Nov 2018, at 01:51, Manan Shah wrote: > > Hello -- thank you so much for your work!...

Thanks for commenting and apologies for the confusion. Only the ‘gcn’ repository was intended to reproduce the results of the paper and hosts the dataset splits that we used (which...

Hi Yifan, Thanks for looking into this. I agree with all of your points. As mentioned previously this data loader is only meant as a 'quick and dirty' example to...

For regression, simply swap out the loss function (L2 instead of cross entropy should do fine). If you have multiple separate graph instances, just concatenate their feature matrices. You can...