smlp
====

.. py:module:: smlp


Classes
-------

.. autoapisummary::

   smlp.SMLP


Module Contents
---------------

.. py:class:: SMLP(input_size=784, hidden_size=77, output_size=6)

   Bases: :py:obj:`torch.nn.Module`


   Simple Multi-Layer Perceptron used in the project.

   The network has four linear layers. The first three are followed by ReLU
   activations and the final layer returns logits. The implementation expects
   flattened inputs of shape ``(batch_size, input_size)`` (for MNIST: input_size=784).

   Parameters
   ----------
   input_size : int
       Dimensionality of the flattened input (default: 784 for 28x28 images).
   hidden_size : int
       Number of neurons in the hidden layers (default: 77 in this project).
   output_size : int
       Number of output classes (default: 6 for digits 4..9 remapped to 0..5).


   .. py:attribute:: layer1


   .. py:attribute:: layer2


   .. py:attribute:: layer3


   .. py:attribute:: layer4


   .. py:attribute:: relu


   .. py:method:: forward(x: torch.Tensor) -> torch.Tensor

      Forward pass.

      Parameters
      ----------
      x : torch.Tensor
          Input tensor of shape (batch_size, input_size). The function does
          not perform flattening; caller must supply a flattened tensor.

      Returns
      -------
      torch.Tensor
          Logits tensor of shape (batch_size, output_size).