dmlp

Classes

DMLP	MLP with 3 hidden layers and 300 neurons in each layer.

Module Contents

class dmlp.DMLP(input_dim, output_dim, hidden_dim=300, negative_slope=0.01)

Bases: torch.nn.Module

MLP with 3 hidden layers and 300 neurons in each layer. LeakyReLU activation function. Args:

input_dim: input dimension output_dim: output dimension hidden_dim: number of neurons in each hidden layer, default 300 negative_slope: negative slope of LeakyReLU activation function, default 0.01

negative_slope = 0.01

net

_init_weights()

He initialization for LeakyReLU

forward(x)

Compute forward pass.

Args:

x: Input tensor of shape (N, input_dim).

Returns: Output tensor of shape (N, output_dim).