pyg_spectral.transforms

class GenNorm(left: float, right: float | None = None, dtype: dtype = torch.float32)

Bases: BaseTransform

Generalized graph normalization.

\[\mathbf{\hat{A}} = \mathbf{\hat{D}}^{-a} (\mathbf{A} + \mathbf{I}) \mathbf{\hat{D}}^{-b}\]

where \(\hat{D}_{ii} = \sum_{j=0} \hat{A}_{ij} + 1\) and \(a,b \in [0,1]\).

Parameters:

• left (float) – left (row) normalization \(a\).

• right (float | None, default: None) – right (col) normalization \(b\). Default to \(1-a\).

forward(data: Data) → Data

_abc_impl = <_abc._abc_data object>

class RemoveSelfLoops(attr: str = 'edge_weight') → None

Bases: BaseTransform

Removes all self-loops in the given homogeneous or heterogeneous graph (functional name: remove_self_loops).

Parameters:

attr (str, optional) – The name of the attribute of edge weights or multi-dimensional edge features to pass to torch_geometric.utils.remove_self_loops(). (default: "edge_weight")

forward(data: Data | HeteroData) → Data | HeteroData

_abc_impl = <_abc._abc_data object>