:mod:`fedlab_core.client.handler`
=================================

.. py:module:: fedlab_core.client.handler


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

Classes
~~~~~~~

.. autoapisummary::

   fedlab_core.client.handler.ClientBackendHandler
   fedlab_core.client.handler.ClientSGDHandler



.. class:: ClientBackendHandler(model, cuda)


   Bases: :class:`object`

   An abstract class representing handler for a client backend.

   In our framework, we define the backend of client handler show manage its local model and buffer.
   It should have a function to update its model called :meth:`train` and a function called :meth:`evaluate`.

   If you use our framework to define the activities of client, please make sure that your self-defined class
   should subclass it. All subclasses should overwrite :meth:`train` and :meth:`evaluate`.

   .. method:: model(self)
      :property:

      Get :class:`torch.nn.Module`


   .. method:: buffer(self)
      :property:

      Get serialized parameters


   .. method:: train(self)
      :abstractmethod:

      Please override this method. This function should manipulate :attr:`self._model` and :attr:`self._buffer`


   .. method:: evaluate(self, test_loader)
      :abstractmethod:

      Please override this method. Evaluate local model based on given test :class:`torch.DataLoader



.. class:: ClientSGDHandler(model, data_loader, optimizer=None, criterion=None, cuda=True, logger_file='log/handler.txt', logger_name='handler')


   Bases: :class:`fedlab_core.client.handler.ClientBackendHandler`

   Client backend handler, this class provides data process method to upper layer.

   :param model:
   :type model: torch.nn.Module
   :param data_loader: :class:`DataLoader` for this client
   :type data_loader: torch.Dataloader
   :param optimizer: optimizer for this client's model. If set to ``None``, will use
   :type optimizer: torch.optim.Optimizer, optional
   :param :func:`torch.optim.SGD` with :attr:`lr` of 0.1 and :attr:`momentum` of 0.9 as default.:
   :param criterion: loss function used in local training process. If set to ``None``, will use
   :type criterion: optional
   :param :func:`nn.CrossEntropyLoss` as default.:
   :param cuda: use GPUs or not. Default: ``True``
   :type cuda: bool, optional
   :param LOGGER: utils class to output debug information to file and command line

   :raises None:

   .. method:: train(self, epochs)

      Client trains its local model on local dataset.

      :param epochs: number of epoch for local training
      :type epochs: int


   .. method:: evaluate(self, test_loader, cuda)

      Evaluate local model based on given test :class:`torch.DataLoader`