trainer.trainer
===============

.. py:module:: trainer.trainer


Attributes
----------

.. autoapisummary::

   trainer.trainer.data_gen


Classes
-------

.. autoapisummary::

   trainer.trainer.AutoTrainer


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

.. py:class:: AutoTrainer(model: Any, loss: torch.nn.Module, optimizer: Any, optimizer_params: Dict = {'lr': 0.0001}, scheduler: Optional[Any] = None, scheduler_params: Optional[Dict] = None, test_eval: Optional[Any] = None)

   Bases: :py:obj:`lightning.LightningModule`


   Class that loads up the configuration for performing training, validation, and testing
   torch.nn.Module models using functionality from PyTorch Lightning.

   Requires lightning.Trainer to perform training and testing.

   If more elaborate settings is needed, please add the necessary configurations
   through inheritence, or directly implement as a subclass of lightning.LightningModule.


   .. attribute:: model

      The model to be trained.

      :type: torch.nn.Module

   .. attribute:: loss

      The loss function used for training.

      :type: function

   .. attribute:: optimizer

      The optimizer used for training.

      :type: type

   .. attribute:: optimizer_params

      Parameters to be passed to the optimizer.

      :type: dict

   .. attribute:: test_eval

      Evaluation function for testing. Defaults to None.

      :type: function

   Initializes an AutoTrainer object.

   :param model: The model to be trained.
   :type model: torch.nn.Module
   :param loss: The loss function used for training.
   :type loss: function
   :param optimizer: The class of the optimizer used for training.
   :type optimizer: type
   :param optimizer_params: Parameters to be passed to the optimizer.
                            Defaults to {"lr": 0.0001}.
   :type optimizer_params: dict
   :param test_eval: Evaluation function for testing. Defaults to None.
   :type test_eval: function, optional


   .. py:attribute:: model


   .. py:attribute:: loss


   .. py:attribute:: optimizer


   .. py:attribute:: optimizer_params


   .. py:attribute:: test_eval
      :value: None



   .. py:attribute:: scheduler
      :value: None



   .. py:attribute:: scheduler_params
      :value: None



   .. py:method:: training_step(batch: Tuple, batch_idx: int) -> torch.Tensor

      Performs a forward pass of the model on a batch of training data
      and returns corresponding loss.

      The computed loss is logged as "train_loss".

      :param batch: The batch used for training.
      :type batch: torch.Tensor
      :param batch_idx: Id corresponding to the used batch.
      :type batch_idx: int

      :returns: training loss from forward pass of batch into model.
      :rtype: torch.Tensor



   .. py:method:: validation_step(batch: Tuple, batch_idx: int) -> None

      Performs a forward pass of the model on a batch of validation data.

      The computed loss is logged as "val_loss".

      :param batch: The batch used for training.
      :type batch: torch.Tensor
      :param batch_idx: Id corresponding to the used batch.
      :type batch_idx: int



   .. py:method:: test_step(batch: Tuple, batch_idx: int) -> None

      Performs a forward pass of the model on a batch of testing data.

      The computed loss is logged as "test_loss".

      :param batch: The batch used for training.
      :type batch: torch.Tensor
      :param batch_idx: Id corresponding to the used batch.
      :type batch_idx: int



   .. py:method:: forward(inputs: Union[torch.Tensor, Any]) -> Union[torch.Tensor, Any]

      Perform one forward pass of the model.

      :param inputs: Input to the model.
      :type inputs: torch.Tensor

      :returns: Model output.
      :rtype: torch.Tensor



   .. py:method:: configure_optimizers() -> Dict[str, Any]

      Define the optimizer of the project.

      :returns: The dictionary containing optimizer and scheduler used for training.
      :rtype: Dict



.. py:data:: data_gen