about code problem

[GorgeousMe]

This is t1_model architecture , but the state load from t1_best_model.pt like this:

so error is :

I don't know how to solve it , could tell me how to do and why ? Very Thanks

[ycjing]

Hi @GorgeousMe

Thank you for your interests in our work. For the problem, I guess it is due to the DGL version. Different DGL versions have different implementations for the GATConv layer. Could you please use the packages indicated at https://github.com/ycjing/AmalgamateGNN.PyTorch/blob/main/requirements.txt and try again? Thanks!

Cheers,
Yongcheng