loki2.models.utils.blocks
=========================

.. py:module:: loki2.models.utils.blocks

.. autoapi-nested-parse::

   Building blocks for neural network architectures.

   This module provides reusable convolutional and deconvolutional blocks
   for use in segmentation networks.





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

.. py:class:: Conv2DBlock(in_channels: int, out_channels: int, kernel_size: int = 3, dropout: float = 0)

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


   Conv2D block with convolution, batch normalization, ReLU, and dropout.

   :param in_channels: Number of input channels for convolution.
   :param out_channels: Number of output channels for convolution.
   :param kernel_size: Kernel size for convolution. Defaults to 3.
   :param dropout: Dropout rate. Defaults to 0.


   .. py:attribute:: block


   .. py:method:: forward(x)


.. py:class:: Deconv2DBlock(in_channels: int, out_channels: int, kernel_size: int = 3, dropout: float = 0)

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


   Deconvolution block with ConvTranspose2d, Conv2d, batch normalization, ReLU, and dropout.

   :param in_channels: Number of input channels for deconv block.
   :param out_channels: Number of output channels for deconv and convolution.
   :param kernel_size: Kernel size for convolution. Defaults to 3.
   :param dropout: Dropout rate. Defaults to 0.


   .. py:attribute:: block


   .. py:method:: forward(x)