dagma.locally_connected.LocallyConnected¶

class dagma.locally_connected.LocallyConnected(num_linear: int, input_features: int, output_features: int, bias: bool = True)¶

Bases: torch.nn.Module

Implements a local linear layer, i.e. Conv1dLocal() with filter size 1.

Parameters:¶

num_linear : int¶

num of local linear layers, i.e.

input_features : int¶

m1

output_features : int¶

m2

bias : bool, optional¶

Whether to include bias or not. Default: True.

weight¶

Type:¶

[d, m1, m2]

bias¶

Type:¶

[d, m2]

Methods¶

reset_parameters()
forward(→ torch.Tensor)	Implements the forward pass of the layer.
extra_repr(→ str)	Returns a string with extra information from the layer.