Learning interpretable latent graphs in the non-Euclidean domain for molecule prediction and healthcare applications (Original code)
We provide two implementations of the suggested GraphInGraph model. In gig_pl is located the implementation used for the second table in * , except DGCNN experiments. In gig_origin can be found our first implementation used to run experiments for the first table. For understanding, the code and experiments we suggest using code from gig_pl, which is clearer and can be much easier to adapt to new datasets. In the future, the code in folder gig_origin will be aligned with code in folder gig_pl.
Precise instructions can be found in each folder in README.md files.