helios.losses.weighted_loss
===========================

.. py:module:: helios.losses.weighted_loss


Classes
-------

.. autoapisummary::

   helios.losses.weighted_loss.WeightedLoss


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

.. py:class:: WeightedLoss(loss_weight: float = 1.0)

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


   Defines a base class for weighted losses.

   The value of the final loss is determined by the following formula:

       .. math:: L_w = w * L

   where :math:`w` is the weight and :math:`L` is the loss function.

   .. rubric:: Example

   .. code-block:: python

       class MyLoss(WeightedLoss):
           ...
           def _eval(self, ...):
               return my_loss_function(...)

   :param loss_weight: the weight of the loss function. Defaults to 1.


   .. py:method:: forward(*args: Any, **kwargs: Any) -> Any

      Forward wrapper function.

      The final loss value will be computed as described above.

      :param \*args: arguments to the loss function.
      :param \*\*kwargs: keyword arguments.

      :returns: The weighted value of the loss function.