eydcao/bsms-gnn

bsms-gnn's Issues

MKL not support M3 chip？

I'm currently using a Macbook Pro shipped with M3 chip. While I try to conda (or pip) install MKL library that is required by sparse-dot-mkl library, it shows no package found error. I suspect it is because M3 chip is not supported by MKL pre-built binary. I want to confirm this issue and would like to ask is there any other alternatives to MKL library such like OpenBLAS?

Layer normalization

Thank you for sharing your good research findings.
I have a question about layer normalization in your network.
Unlike conventional cnn or transformer, one layer normalization is applied after going through several gcn.
Is there a special reason for applying it at the end of the encoder instead of applying it after each gcn layer?

Could you provide the implementation of MS-GNN-GRID baseline?

Deforming plate

Hi, my work is related to solids, i've tried MeshGraphNets now i want to assess the speed of your implementation,
so far i've resolved some of the issues like the dependency ones but there are a few things that i need to ask, or maybe more if model won't run (i'm using code from ur previous implementation). I would really appreciate if you'd take some time to answer my queries.
Thanks @Eydcao

Could you please give exact package version in the requirements?

Transition between levels

Dear Yadi Cao,

I have a question about your model.

When the transition is happening between the different layers, does the node samples itself into the next layer or only the neighbouring nodes are aggregated? With other words, do self edges exist is the convolutional step between layers?

Thank you four your kind help!