Changelog
=========

Version 2.4.0 (2024-10-02)
---------------------------

**New Features**

- Added ``strategy="fsdp_native"`` for PyTorch 2.0+ native FSDP with activation checkpointing.
- New :class:`~lightning.pytorch.callbacks.StochasticWeightAveraging` callback.
- Added ``Trainer(barebones=True)`` mode — disables all non-essential features for maximum throughput.
- ``LightningModule.log_dict()`` now supports nested dict keys (e.g., ``"metrics/train/loss"``).
- New :class:`~lightning.pytorch.loggers.NeptuneLogger` for Neptune.ai integration.
- ``Trainer(compile=True)`` shorthand to compile the model with ``torch.compile()``.
- Async checkpointing: checkpoints are now saved in a background thread by default.

**Breaking Changes**

- Dropped Python 3.8 support. Minimum is now Python 3.9.
- ``Trainer(gpus=...)`` and ``Trainer(tpu_cores=...)`` removed (deprecated since 2.0). Use ``devices=`` instead.
- ``LightningModule.on_post_move_to_device()`` removed.

**Deprecations**

- ``Trainer(auto_lr_find=True)`` deprecated; use ``from lightning.pytorch.tuner import Tuner`` instead.

Version 2.3.0 (2024-07-15)
---------------------------

**New Features**

- Full PyTorch 2.3 compatibility including ``torch.compile()`` integration.
- New Lightning CLI — train models from YAML config files without writing Python:

  .. code-block:: bash

     lightning fit --config config.yaml
     lightning validate --config config.yaml --ckpt_path best
     lightning predict --config config.yaml --ckpt_path last

**Bug Fixes**

- Fixed gradient accumulation with ``accumulate_grad_batches > 1`` and ``strategy="fsdp"``.
- Fixed ``ModelCheckpoint.best_model_path`` returning ``None`` after ``trainer.test()``.
- Fixed memory leak in ``RichProgressBar`` when ``leave=False``.

Version 2.2.0 (2024-04-11)
---------------------------

**New Features**

- ``LightningModule`` now supports ``configure_model()`` hook for model parallel setups.
- Added ``Trainer(reload_dataloaders_every_n_epochs=N)`` parameter.
- New ``lightning.pytorch.utilities.grad_norm`` utility function.

Version 2.0.0 (2023-09-01)
---------------------------

**Major Release**

- Complete rewrite of the Trainer internals.
- Minimum PyTorch bumped to 2.0.
- Introduced ``lightning.fabric.Fabric`` — a lightweight alternative to ``Trainer`` for fine-grained control.
- New plugin system for custom accelerators and strategies.
- ``DeepSpeed`` strategy updated to DeepSpeed 0.12.
- Unified ``lightning`` package combining ``pytorch-lightning`` and ``lightning-fabric``.