:py:mod:`kosmos.ml.config.factories.optimizer`
==============================================

.. py:module:: kosmos.ml.config.factories.optimizer






Module Attributes
-----------------

.. py:type:: ParamsT
   :canonical: Iterable[torch.Tensor] | Iterable[dict[str, Any]] | Iterable[tuple[str, torch.Tensor]]





Classes
-------

.. py:class:: OptimizerConfig

   Bases: :py:class:`abc.ABC`

   Optimizer configuration.





   |

   .. rubric:: Methods

   .. py:method:: get_instance(params: ParamsT) -> torch.optim.Optimizer

      Get the optimizer instance.

      :param params: Parameters to optimize.
      :type params: ParamsT

      :returns: Optimizer instance.
      :rtype: Optimizer




----

.. py:class:: SGDOptimizerConfig(lr: float = 0.001, momentum: float = 0.0, weight_decay: float = 0.0, *, nesterov: bool = False)

   Bases: :py:class:`OptimizerConfig`

   Stochastic gradient descent (SGD) optimizer configuration.

   Initialize the SGD optimizer configuration.

   :param lr: Learning rate. Defaults to 1e-3.
   :type lr: float
   :param momentum: Momentum factor. Defaults to 0.0.
   :type momentum: float
   :param weight_decay: Weight decay. Defaults to 0.0.
   :type weight_decay: float
   :param nesterov: Whether to use Nesterov momentum. Only applicable
                    when momentum is non-zero. Defaults to False.
   :type nesterov: bool





   |

   .. rubric:: Methods

   .. py:method:: get_instance(params: ParamsT) -> torch.optim.SGD

      Get the SGD optimizer instance.

      :param params: Parameters to optimize.
      :type params: ParamsT

      :returns: SGD optimizer instance.
      :rtype: SGD




----

.. py:class:: AdamOptimizerConfig(lr: float = 0.001, weight_decay: float = 0.0)

   Bases: :py:class:`OptimizerConfig`

   Adam optimizer configuration.

   Initialize the Adam optimizer configuration.

   :param lr: Learning rate. Defaults to 1e-3.
   :type lr: float
   :param weight_decay: Weight decay. Defaults to 0.0.
   :type weight_decay: float





   |

   .. rubric:: Methods

   .. py:method:: get_instance(params: ParamsT) -> torch.optim.Adam

      Get the Adam optimizer instance.

      :param params: Parameters to optimize.
      :type params: ParamsT

      :returns: Adam optimizer instance.
      :rtype: Adam