gammagl.models.Specformer

class Specformer(nclass, n_feat, n_layer=2, hidden_dim=32, n_heads=4, tran_dropout=0.2, feat_dropout=0.4, prop_dropout=0.5, norm='none')

The Specformer from the “Specformer:Spectral Graph Neural Networks Meet Transformers” paper

Parameters:

• nclass (int) – the number of node classes

• n_feat (int) – the node feature input dimension

• n_layer (int) – number of Speclayers

• hidden_dim (int) – the eigvalue representation dimension and the node feature dimension

• n_heads (int) – the number of attention heads

• tran_dropout (int) – the probability of dropout

• feat_dropout (int) – the probability of dropout

• prop_dropout (float) – the probability of dropout

forward(x, edge_index, e, u)

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.