shredx.modules.mlp.MLPEncoder

class shredx.modules.mlp.MLPEncoder(input_size: int, hidden_size: int, num_layers: int, dropout: float, device: str = 'cpu')

Bases: Module

Multi-layer perceptron (MLP) encoder.

Creates a feed-forward neural network with identical layer sizes. Uses ReLU activations between layers and applies dropout after the final layer.

Parameters:

input_sizeint

Input feature dimension.

hidden_sizeint

Hidden state dimension.

num_layersint

Number of stacked MLP layers.

dropoutfloat

Dropout probability applied to the outputs.

devicestr, optional

Device on which to place the module. Default is "cpu".

Methods

forward(x)

Apply the MLP encoder to an input batch.

Notes

Class Methods:

forward(x):

• Applies the MLP encoder to an input batch.

• Parameters:

  • x : Float[torch.Tensor, "batch sequence input_size"]. Input tensor.

• Returns:

  • tuple. Tuple containing the final output tensor of shape (batch_size, 1, 1, hidden_size) and None for no auxiliary losses.