Contents:
GraphEncoder.
forward
Encode graph to vertex latent distribution
eidx (Tensor) – Vertex indices of edges (\(2 \times n_{edges}\))
Tensor
enorm (Tensor) – Normalized weight of edges (\(n_{edges}\))
esgn (Tensor) – Sign of edges (\(n_{edges}\))
Normal
v – Vertex latent distribution (\(n_{vertices} \times n_{features}\))