Testing PL Module. Use as follows: - subclass - modify the behavior for what you want .. warning:: This is meant for testing/debugg

r"""Fabric accelerates your PyTorch training or inference code with minimal changes required. Key Features: - Automatic placement of mode

r"""Save the model after every epoch by monitoring a quantity. Every logged metrics are passed to the :class:`~lightning.pytorch.loggers.logger.Lo

.. warning:: This is meant for testing/debugging and is experimental.

Exception used to inform users of misuse with Lightning.

The Connector parses several Fabric arguments and instantiates the Strategy including its owned components. A. accelerator flag could be:

Implementation of a configurable command line tool for pytorch-lightning.

r"""Strategy for Fully Sharded Data Parallel provided by torch.distributed. Fully Sharded Training shards the entire model across all available G

r"""Log to the local file system in CSV format. Logs are saved to ``os.path.join(root_dir, name, version)``. Args: root_dir: The roo

Enables user-defined parallelism applied to a model. .. warning:: This is an :ref:`experimental <versioning:Experimental API>` feature. Cur

r"""Log to local file system in `TensorBoard <https://www.tensorflow.org/tensorboard>`_ format. Implemented using :class:`~tensorboardX.SummaryWr

Tuner class to tune your model.

Strategy for multi-process single-device training on one or multiple nodes.

Combines different iterables under specific sampling modes. Args: iterables: the iterable or collection of iterables to sample from.

Logger that enables remote experiment tracking, logging, and artifact management on lightning.ai.

r"""Monitor a metric and stop training when it stops improving. Args: monitor: quantity to be monitored. min_delta: minimum chang

Strategy that handles communication on a single device.

Precision plugin for training with Fully Sharded Data Parallel (FSDP). .. warning:: This is an :ref:`experimental <versioning:Experimental API>`

r"""Automatically monitor and logs learning rate for learning rate schedulers during training. Args: logging_interval: set to ``'epoch'``

The default environment used by Lightning for a single node or free cluster (not managed). There are two modes the Lightning environment can oper

Create a progress bar with `rich text formatting <https://github.com/Textualize/rich>`_. Install it with pip: .. code-block:: bash

r"""This is the default progress bar used by Lightning. It prints to ``stdout`` using the :mod:`tqdm` package and shows up to four different bars:

A context manager to change the default tensor type when tensors get created. See: :func:`torch.set_default_dtype`

r"""Launches processes that run a given function in parallel, and joins them all at the end. The main process in which this launcher is invoked c

.. warning:: This is meant for testing/debugging and is experimental.

Log using `MLflow <https://mlflow.org>`_. Install it with pip: .. code-block:: bash pip install mlflow # or mlflow-skinny ..

Plugin for Automatic Mixed Precision (AMP) training with ``torch.autocast``. Args: precision: Whether to use ``torch.float16`` (``'16-mix

Plugin for training with XLA. Args: precision: Full precision (32-true) or half precision (16-true, bf16-true). Raises: Valu

Cluster environment for training on a cluster managed by SLURM. You can configure the `main_address` and `main_port` properties via the env varia

r"""Log using `Weights and Biases <https://docs.wandb.ai/guides/integrations/lightning>`_. **Installation and set-up** Install with pip:

Strategy for training multiple TPU devices using the :func:`torch_xla.distributed.xla_multiprocessing.spawn` method.

Base class for all plugins handling the precision-specific parts of the training. The class attribute precision must be overwritten in child clas

Precision plugin for DeepSpeed integration. Args: precision: Full precision (32-true), half precision (16-true, bf16-true) or

Plugin for training with half precision. Args: precision: Whether to use ``torch.float16`` (``'16-true'``) or ``torch.bfloat16`` (``'bf16

r"""Generates a summary of all layers in a :class:`~lightning.pytorch.core.LightningModule`. Args: max_depth: The maximum depth of layer

The Timer callback tracks the time spent in the training, validation, and test loops and interrupts the Trainer if the given time limit for the tr

r"""A process launcher that invokes the current script as many times as desired in a single node. This launcher needs to be invoked on each node.

Customized `MNIST <http://yann.lecun.com/exdb/mnist/>`_ dataset for testing PyTorch Lightning without the torchvision dependency. Part of the

Collection (dictionary) of :class:`~lightning.pytorch.trainer.connectors.logger_connector.result._ResultMetric` Example:: # you can log

Registers legacy artifacts (classes, methods, etc.) that were removed but still need to be included for unpickling old checkpoints. The following

This profiler simply records the duration of actions (in seconds) and reports the mean duration of each action and the total time spent over the e

r"""Computes and logs throughput with the :class:`~lightning.fabric.utilities.throughput.Throughput` Example:: class MyModel(LightningMo

Cluster environment for training on a TPU Pod with the `PyTorch/XLA <https://pytorch.org/xla>`_ library. A list of environment variables set by X

r"""Track your parameters, metrics, source code and more using `Comet <https://www.comet.com/?utm_source=lightning.pytorch&utm_medium=referral>`_.

r"""Change gradient accumulation factor according to scheduling. Args: scheduling: scheduling in format {epoch: accumulation_factor}

.. warning:: This is meant for testing/debugging and is experimental.

Mini version of WikiText2.

Accelerator for XLA devices, normally TPUs. .. warning:: Use of this accelerator beyond import and instantiation is experimental.

r"""Strategy for training multiple XLA devices using the :func:`torch_xla.distributed.xla_fully_sharded_data_parallel.XlaFullyShardedDataParallel`

A container to store state variables of your program. This is a drop-in replacement for a Python dictionary, with the additional functionality to

Finds the largest batch size supported by a given model before encountering an out of memory (OOM) error. All you need to do is add it as a callb

Plugin for quantizing weights with `bitsandbytes <https://github.com/bitsandbytes-foundation/bitsandbytes>`__. .. warning:: This is an :ref:`exp

Exponential Moving Average (EMA) Weight Averaging callback.

Captures the ``nn.Module.training`` (bool) mode of every submodule, and allows it to be restored later on.

Track aggregated and current progress. Args: total: Intended to track the total progress of an event. current: Intended to track

Wraps the value provided to `:meth:`~lightning.pytorch.core.LightningModule.log`

r"""Automatically monitors and logs device stats during training, validation and testing stage. ``DeviceStatsMonitor`` is a special callback as it

Plugin for training with double (``torch.float64``) precision.

Environment for distributed training using the `PyTorchJob`_ operator from `Kubeflow`_. This environment, unlike others, does not get auto-detect

An environment for running on clusters managed by the LSF resource manager. It is expected that any execution using this ClusterEnvironment was e

The ``LearningRateFinder`` callback enables the user to do a range test of good initial learning rates, to reduce the amount of guesswork in picki

Used to save a checkpoint on exception. Args: dirpath: directory to save the checkpoint file. filename: checkpoint filename. This

r"""Generates a summary of all layers in a :class:`~lightning.pytorch.core.LightningModule` with `rich text formatting <https://github.com/Textual

CheckpointIO that utilizes :func:`torch.save` and :func:`torch.load` to save and load checkpoints respectively, common for most use cases. ..