This repository implements learning a Sheaf Laplacian within a kernel function of a Gaussian process, giving us a Sheaf GP.
The main implementation using the same splits as Transductive Kernels for Gaussian Processes on Graphs can be found in sheaf_gp.py
.
Implementation using the same splits as Neural Sheaf Diffusion can be found in sheaf_gp_nsd.py
, the raw files are stored in the splits
folder.
Lastly, various kernels can be found in kernels.py
.