Just wondering whether you would be willing to share any insights into how to best adapt the codebase to a new 3D graph dataset. We have a dataset of 3D building wireframes, where each building wireframe is represented as a list of 3D points and the corresponding adjacency matrix.
Would you recommend building on top of GruM-2D or GruM-3D? What would be your suggested approach?
Intuitively, I would think GruM-3D, but GruM-3D seems heavily geared towards molecule datasets and your README states that GruM-2D is better suited for general graph modeling tasks.
Hi,
Thank you for the great work!
Just wondering whether you would be willing to share any insights into how to best adapt the codebase to a new 3D graph dataset. We have a dataset of 3D building wireframes, where each building wireframe is represented as a list of 3D points and the corresponding adjacency matrix.
Would you recommend building on top of GruM-2D or GruM-3D? What would be your suggested approach?
Intuitively, I would think GruM-3D, but GruM-3D seems heavily geared towards molecule datasets and your README states that GruM-2D is better suited for general graph modeling tasks.
Thank you, Kevin