torchcell.models.hetero_cell_bipartite_dango_gi_lazy.PreProcessor

class PreProcessor(in_channels: int, hidden_channels: int, num_layers: int = 2, dropout: float = 0.1, norm: str = 'layer', activation: str = 'relu')

Bases: Module

MLP that preprocesses node features with linear, norm, activation, and dropout layers.

forward(x: Tensor) → Tensor

Pass node features through the preprocessing MLP.