Mufei Li
Are you following a particular tutorial?
I tried it myself and the program was killed. I guess ogbn-products is currently too large for gnnlens to plot at once. Could you try extracting a subgraph and then...
I suspect the issue is related to [L325](https://github.com/rajarshiche/GNNs/blob/main/GAT_trial1.py#L325). If you call `dgl.add_self_loop(bg)`, the new graph returned will have batch size 1. As a result, the first dimension of the model...
Thanks for the update and I'm glad you solved that.
Sure. I'll check it later. Thanks.
While I agree this name can be misleading, directly renaming it might break too much backward compatibility. One possible workaround is to add an alias of the class `MoleculeDFDataset`.
What's the version for `dgl` and `dgllife`? I tried your code snippet on a different protein pdb file, which seems to be working fine. > Also do pos1 and pos2...
Why did you do the following? ```bash batch = dgl.graph([g1, g2]) ``` If you want to batch two graphs, you need to do ```bash batch = dgl.batch([g1, g2]) ```
Can you share `4wtg.pdb` for reproducing the issue?
I think the code snippet does not involve passing the input to the model. Can you also include those lines of code?