---
title: Upgrade from 1.4 to the 2.0
canonical: 'https://lightning.ai/docs/pytorch/stable/home/migration_guide/from_1_4'
---
# Upgrade from 1.4 to the 2.0 {#upgrade-from-1-4-to-the-2-0}

## Regular User {#regular-user}

| If | Then | Ref |
| --- | --- | --- |
| relied on the `outputs` in your  `LightningModule.on_train_epoch_end` or `Callback.on_train_epoch_end` hooks | rely on either `on_train_epoch_end` or set outputs as attributes in your `LightningModule` instances and access them from the hook | [PR7339](https://github.com/Lightning-AI/pytorch-lightning/pull/7339) |
| accessed `Trainer.truncated_bptt_steps` | switch to manual optimization | [PR7323](https://github.com/Lightning-AI/pytorch-lightning/pull/7323) |
| called  `LightningModule.write_predictions`  and  `LightningModule.write_predictions_dict` | rely on `predict_step` and `Trainer.predict` + callbacks to write out predictions | [PR7066](https://github.com/Lightning-AI/pytorch-lightning/pull/7066) |
| passed the `period` argument to the `ModelCheckpoint` callback | pass the `every_n_epochs` argument to the `ModelCheckpoint` callback | [PR6146](https://github.com/Lightning-AI/pytorch-lightning/pull/6146) |
| passed the `output_filename` argument to `Profiler` | now pass `dirpath` and `filename`, that is  `Profiler(dirpath=...., filename=...)` | [PR6621](https://github.com/Lightning-AI/pytorch-lightning/pull/6621) |
| passed the `profiled_functions` argument in  `PytorchProfiler` | now pass the  `record_functions` argument | [PR6349](https://github.com/Lightning-AI/pytorch-lightning/pull/6349) |
| relied on the `@auto_move_data` decorator to use the `LightningModule` outside of the `Trainer` for inference | use `Trainer.predict` | [PR6993](https://github.com/Lightning-AI/pytorch-lightning/pull/6993) |
| implemented `on_load_checkpoint` with a `checkpoint` only argument, as in `Callback.on_load_checkpoint(checkpoint)` | now update the signature to include `pl_module` and `trainer`, as in `Callback.on_load_checkpoint(trainer, pl_module, checkpoint)` | [PR7253](https://github.com/Lightning-AI/pytorch-lightning/pull/7253) |
| relied on `pl.metrics` | now import separate package `torchmetrics` | [torchmetrics](https://torchmetrics.readthedocs.io/en/stable) |
| accessed `datamodule` attribute of `LightningModule`, that is `model.datamodule` | now access `Trainer.datamodule`, that is `model.trainer.datamodule` | [PR7168](https://github.com/Lightning-AI/pytorch-lightning/pull/7168) |

| If | Then | Ref |
| --- | --- | --- |
| used `trainer.fit(train_dataloaders=...)` | use `trainer.fit(dataloaders=...)` | [PR7431](https://github.com/Lightning-AI/pytorch-lightning/pull/7431) |
| used `trainer.validate(val_dataloaders...)` | use  `trainer.validate(dataloaders=...)` | [PR7431](https://github.com/Lightning-AI/pytorch-lightning/pull/7431) |
| passed `num_nodes`  to  `DDPPlugin` and `DDPSpawnPlugin` | remove them since these parameters are now passed from the `Trainer` | [PR7026](https://github.com/Lightning-AI/pytorch-lightning/pull/7026) |
| passed `sync_batchnorm` to `DDPPlugin` and `DDPSpawnPlugin` | remove them since these parameters are now passed from the `Trainer` | [PR7026](https://github.com/Lightning-AI/pytorch-lightning/pull/7026) |
| didn’t provide a `monitor` argument to the `EarlyStopping` callback and just relied on the default value | pass  `monitor` as it is now a required argument | [PR7907](https://github.com/Lightning-AI/pytorch-lightning/pull/7907) |
| used `every_n_val_epochs` in `ModelCheckpoint` | change the argument to `every_n_epochs` | [PR8383](https://github.com/Lightning-AI/pytorch-lightning/pull/8383) |
| used Trainer’s flag `reload_dataloaders_every_epoch` | use pass `reload_dataloaders_every_n_epochs` | [PR5043](https://github.com/Lightning-AI/pytorch-lightning/pull/5043) |
| used Trainer’s flag `distributed_backend` | use `strategy` | [PR8575](https://github.com/Lightning-AI/pytorch-lightning/pull/8575) |

| If | Then | Ref |
| --- | --- | --- |
| used Trainer’s flag `terminate_on_nan` | set `detect_anomaly` instead, which enables detecting anomalies in the autograd engine | [PR9175](https://github.com/Lightning-AI/pytorch-lightning/pull/9175) |
| used Trainer’s flag `weights_summary` | pass a `ModelSummary` callback with `max_depth` instead | [PR9699](https://github.com/Lightning-AI/pytorch-lightning/pull/9699) |
| used Trainer’s flag `checkpoint_callback` | set `enable_checkpointing`. If you set `enable_checkpointing=True`, it configures a default `ModelCheckpoint` callback if none is provided `lightning.pytorch.trainer.trainer.Trainer.callbacks.ModelCheckpoint` | [PR9754](https://github.com/Lightning-AI/pytorch-lightning/pull/9754) |
| used Trainer’s flag `stochastic_weight_avg` | add the `StochasticWeightAveraging` callback directly to the list of callbacks, so for example, `Trainer(..., callbacks=[StochasticWeightAveraging(), ...])` | [PR8989](https://github.com/Lightning-AI/pytorch-lightning/pull/8989) |
| used Trainer’s flag `flush_logs_every_n_steps` | pass it to the logger init if it is supported for the particular logger | [PR9366](https://github.com/Lightning-AI/pytorch-lightning/pull/9366) |
| used Trainer’s flag `max_steps` to the `Trainer`, `max_steps=None` won't have any effect | turn off the limit by passing `Trainer(max_steps=-1)` which is the default | [PR9460](https://github.com/Lightning-AI/pytorch-lightning/pull/9460) |
| used Trainer’s flag `resume_from_checkpoint="..."` | pass the same path to the fit function instead, `trainer.fit(ckpt_path="...")` | [PR9693](https://github.com/Lightning-AI/pytorch-lightning/pull/9693) |
| used Trainer’s flag `log_gpu_memory`, `gpu_metrics` | use the `DeviceStatsMonitor` callback instead | [PR9921](https://github.com/Lightning-AI/pytorch-lightning/pull/9921) |
| used Trainer’s flag `progress_bar_refresh_rate` | set the `ProgressBar` callback and set `refresh_rate` there, or pass `enable_progress_bar=False` to disable the progress bar | [PR9616](https://github.com/Lightning-AI/pytorch-lightning/pull/9616) |
| called `LightningModule.summarize()` | use the utility function `pl.utilities.model_summary.summarize(model)` | [PR8513](https://github.com/Lightning-AI/pytorch-lightning/pull/8513) |
| used the `LightningModule.model_size` property | use the utility function `pl.utilities.memory.get_model_size_mb(model)` | [PR8495](https://github.com/Lightning-AI/pytorch-lightning/pull/8495) |
| relied on the `on_train_dataloader()` hooks in  `LightningModule` and `LightningDataModule` | use `train_dataloader` | [PR9098](https://github.com/Lightning-AI/pytorch-lightning/pull/9098) |
| relied on the `on_val_dataloader()` hooks in  `LightningModule` and `LightningDataModule` | use `val_dataloader` | [PR9098](https://github.com/Lightning-AI/pytorch-lightning/pull/9098) |
| relied on the `on_test_dataloader()` hooks in  `LightningModule` and `LightningDataModule` | use `test_dataloader` | [PR9098](https://github.com/Lightning-AI/pytorch-lightning/pull/9098) |
| relied on the `on_predict_dataloader()` hooks in  `LightningModule` and `LightningDataModule` | use `predict_dataloader` | [PR9098](https://github.com/Lightning-AI/pytorch-lightning/pull/9098) |
| implemented the `on_keyboard_interrupt` callback hook | implement the `on_exception` hook, and specify the exception type | [PR9260](https://github.com/Lightning-AI/pytorch-lightning/pull/9260) |
| relied on the `TestTubeLogger` | Use another logger like `TensorBoardLogger` | [PR9065](https://github.com/Lightning-AI/pytorch-lightning/pull/9065) |
| used the basic progress bar `ProgressBar` callback | use the `TQDMProgressBar` callback instead with the same arguments | [PR10134](https://github.com/Lightning-AI/pytorch-lightning/pull/10134) |
| were using `GPUStatsMonitor`  callbacks | use `DeviceStatsMonitor` callback instead | [PR9924](https://github.com/Lightning-AI/pytorch-lightning/pull/9924) |
| were using `XLAStatsMonitor` callbacks | use `DeviceStatsMonitor` callback instead | [PR9924](https://github.com/Lightning-AI/pytorch-lightning/pull/9924) |

| If | Then | Ref |
| --- | --- | --- |
| have wrapped your loggers with `LoggerCollection` | directly pass a list of loggers to the Trainer and access the list via the `trainer.loggers` attribute. | [PR12147](https://github.com/Lightning-AI/pytorch-lightning/pull/12147) |
| used `Trainer.lr_schedulers` | access `trainer.lr_scheduler_configs` instead, which contains dataclasses instead of dictionaries. | [PR11443](https://github.com/Lightning-AI/pytorch-lightning/pull/11443) |
| used `neptune-client` API in the `NeptuneLogger` | upgrade to the latest API | [PR14727](https://github.com/Lightning-AI/pytorch-lightning/pull/14727) |
| used  `LightningDataModule.on_save` hook | use  `LightningDataModule.on_save_checkpoint` instead | [PR11887](https://github.com/Lightning-AI/pytorch-lightning/pull/11887) |
| used  `LightningDataModule.on_load_checkpoint` hook | use  `LightningDataModule.on_load_checkpoint` hook instead | [PR11887](https://github.com/Lightning-AI/pytorch-lightning/pull/11887) |
| used  `LightningModule.on_hpc_load` hook | switch to general purpose hook `LightningModule.on_load_checkpoint` | [PR14315](https://github.com/Lightning-AI/pytorch-lightning/pull/14315) |
| used  `LightningModule.on_hpc_save` hook | switch to general purpose hook `LightningModule.on_save_checkpoint` | [PR14315](https://github.com/Lightning-AI/pytorch-lightning/pull/14315) |
| used Trainer’s flag `weights_save_path` | use directly `dirpath` argument in the `ModelCheckpoint` callback. | [PR14424](https://github.com/Lightning-AI/pytorch-lightning/pull/14424) |
| used Trainer’s property `Trainer.weights_save_path` is dropped |  | [PR14424](https://github.com/Lightning-AI/pytorch-lightning/pull/14424) |

| If | Then | Ref |
| --- | --- | --- |
| used `seed_everything_default=None` in `LightningCLI` | set `seed_everything_default=False` instead | [PR12804](https://github.com/Lightning-AI/pytorch-lightning/pull/12804) |
| used `Trainer.reset_train_val_dataloaders()` | call `Trainer.fit_loop.setup_data()` | [PR12184](https://github.com/Lightning-AI/pytorch-lightning/pull/12184) |
| imported `pl.core.lightning` | import `pl.core.module` instead | [PR12740](https://github.com/Lightning-AI/pytorch-lightning/pull/12740) |

| If | Then | Ref |
| --- | --- | --- |
| used Python 3.7 | upgrade to Python 3.8 or higher | [PR16579](https://github.com/Lightning-AI/pytorch-lightning/pull/16579) |
| used PyTorch 1.10 | upgrade to PyTorch 1.11 or higher | [PR16492](https://github.com/Lightning-AI/pytorch-lightning/pull/16492) |
| used Trainer’s flag `gpus` | use `devices` with the same number | [PR16171](https://github.com/Lightning-AI/pytorch-lightning/pull/16171) |
| used Trainer’s flag `tpu_cores` | use `devices` with the same number | [PR16171](https://github.com/Lightning-AI/pytorch-lightning/pull/16171) |
| used Trainer’s flag `ipus` | use `devices` with the same number | [PR16171](https://github.com/Lightning-AI/pytorch-lightning/pull/16171) |
| used Trainer’s flag `num_processes` | use `devices` with the same number | [PR16171](https://github.com/Lightning-AI/pytorch-lightning/pull/16171) |
| used Trainer’s flag `resume_from_checkpoint` | pass the path to the `Trainer.fit(ckpt_path="...")` method, | [PR10061](https://github.com/Lightning-AI/pytorch-lightning/pull/10061) |
| used Trainer’s flag `auto_select_gpus` | use `devices="auto"` | [PR16184](https://github.com/Lightning-AI/pytorch-lightning/pull/16184) |
| called the `pl.tuner.auto_gpu_select.pick_single_gpu` function | use Trainer’s flag `devices="auto"` | [PR16184](https://github.com/Lightning-AI/pytorch-lightning/pull/16184) |
| called the `pl.tuner.auto_gpu_select.pick_multiple_gpus` functions | use Trainer’s flag `devices="auto"` | [PR16184](https://github.com/Lightning-AI/pytorch-lightning/pull/16184) |
| used Trainer’s flag  `accumulate_grad_batches` with a scheduling dictionary value | use the  `GradientAccumulationScheduler` callback and configure it | [PR16729](https://github.com/Lightning-AI/pytorch-lightning/pull/16729) |
| imported profiles from `pl.profiler` | import from `pl.profilers` | [PR16359](https://github.com/Lightning-AI/pytorch-lightning/pull/16359) |
| used `Tuner` as part of `Trainer` in any form | move to a standalone `Tuner` object or use particular callbacks `LearningRateFinder` and `BatchSizeFinder` | [batch_size_finder](/docs/pytorch/2.6.6/reference/advanced/training_tricks.md) [learning_rate_finder](/docs/pytorch/2.6.6/reference/advanced/training_tricks.md) |
| used Trainer’s flag `auto_scale_batch_size` | use `BatchSizeFinder` callback instead and the `Trainer.tune()` method was removed |  |
| used Trainer’s flag `auto_lr_find` | use callbacks `LearningRateFinder` callback instead and the `Trainer.tune()` method was removed |  |

## Advanced User {#advanced-user}

| If | Then | Ref |
| --- | --- | --- |
| called `ModelCheckpoint.save_function` | now call `Trainer.save_checkpoint` | [PR7201](https://github.com/Lightning-AI/pytorch-lightning/pull/7201) |
| accessed the `Trainer.running_sanity_check` property | now  access the `Trainer.sanity_checking` property | [PR4945](https://github.com/Lightning-AI/pytorch-lightning/pull/4945) |
| used `LightningModule.grad_norm` | now use the `pl.utilities.grad_norm` utility function instead | [PR7292](https://github.com/Lightning-AI/pytorch-lightning/pull/7292) |
| used `TrainerTrainingTricksMixin.detect_nan_tensors` | now use `pl.utilities.grads.grad_norm` | [PR6834](https://github.com/Lightning-AI/pytorch-lightning/pull/6834) |
| used `TrainerTrainingTricksMixin.print_nan_gradients` | now use `pl.utilities.finite_checks.print_nan_gradients` | [PR6834](https://github.com/Lightning-AI/pytorch-lightning/pull/6834) |
| If you relied on `TrainerLoggingMixin.metrics_to_scalars` | now use `pl.utilities.metrics.metrics_to_scalars` | [PR7180](https://github.com/Lightning-AI/pytorch-lightning/pull/7180) |
| selected the i-th GPU with `Trainer(gpus="i,j")` | now this will set the number of GPUs, just like passing `Trainer(devices=i)`, you can still select the specific GPU by setting the `CUDA_VISIBLE_DEVICES=i,j` environment variable | [PR6388](https://github.com/Lightning-AI/pytorch-lightning/pull/6388) |

| If | Then | Ref |
| --- | --- | --- |
| used `self.log(sync_dist_op=...)` | use `self.log(reduce_fx=...)` instead. Passing `"mean"` will still work, but it also takes a callable | [PR7891](https://github.com/Lightning-AI/pytorch-lightning/pull/7891) |
| used the argument `model` from `pytorch_lightning.utilities.model_helper.is_overridden` | use `instance` instead | [PR7918](https://github.com/Lightning-AI/pytorch-lightning/pull/7918) |
| returned values from `training_step` that had `.grad` defined (e.g., a loss) and expected `.detach()` to be called for you | call `.detach()` manually | [PR7994](https://github.com/Lightning-AI/pytorch-lightning/pull/7994) |
| imported `pl.utilities.distributed.rank_zero_warn` | import `pl.utilities.rank_zero.rank_zero_warn` |  |
| relied on `DataModule.has_prepared_data` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_setup_fit` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_setup_validate` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_setup_test` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_setup_predict` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_teardown_fit` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_teardown_validate` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_teardown_test` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| relied on `DataModule.has_teardown_predict` attribute | manage data lifecycle in customer methods | [PR7657](https://github.com/Lightning-AI/pytorch-lightning/pull/7657) |
| used `DDPPlugin.task_idx` | use `DDPStrategy.local_rank` | [PR8203](https://github.com/Lightning-AI/pytorch-lightning/pull/8203) |
| used `Trainer.disable_validation` | use the condition `not Trainer.enable_validation` | [PR8291](https://github.com/Lightning-AI/pytorch-lightning/pull/8291) |

| If | Then | Ref |
| --- | --- | --- |
| passed `prepare_data_per_node` to the `Trainer` | set it as a property of `DataHooks`, accessible in the `LightningModule` and `LightningDataModule` instead | [PR8958](https://github.com/Lightning-AI/pytorch-lightning/pull/8958) |
| used  `process_position` flag | specify your  `ProgressBar` callback and set it as `process_position` directly | [PR9222](https://github.com/Lightning-AI/pytorch-lightning/pull/9222) |
| used distributed training attributes `add_to_queue` and `get_from_queue` in `LightningModule` | user the same methods in `DDPStrategy(start_method='spawn')` | [PR9118](https://github.com/Lightning-AI/pytorch-lightning/pull/9118) |
| called `LightningModule.get_progress_bar_dict` | use the utility function `pl.callbacks.progress.base.get_standard_metrics(module.trainer)` | [PR9118](https://github.com/Lightning-AI/pytorch-lightning/pull/9118) |
| used `LightningModule.on_post_move_to_device` | remove it as parameters tying happens automatically without the need of implementing your own logic | [PR9525](https://github.com/Lightning-AI/pytorch-lightning/pull/9525) |
| relied on  `Trainer.progress_bar_dict` | use  `ProgressBarBase.get_metrics` | [PR9118](https://github.com/Lightning-AI/pytorch-lightning/pull/9118) |
| used `LightningDistributed` | rely on the logic in `DDPStrategy(start_method='...')` | [PR9691](https://github.com/Lightning-AI/pytorch-lightning/pull/9691) |
| used the Accelerator collective API `Accelerator.barrier`, `Accelerator.broadcast`, and `Accelerator.all_gather` | call `Strategy` collectives API directly, without going through `Accelerator` | [PR9677](https://github.com/Lightning-AI/pytorch-lightning/pull/9677) |
| used `pytorch_lightning.core.decorators.parameter_validation` | rely on automatic parameters tying with `pytorch_lightning.utilities.params_tying.set_shared_parameters` | [PR9525](https://github.com/Lightning-AI/pytorch-lightning/pull/9525) |
| used `LearningRateMonitor.lr_sch_names` | access them using `LearningRateMonitor.lrs.keys()` which will return the names of all the optimizers, even those without a scheduler. | [PR10066](https://github.com/Lightning-AI/pytorch-lightning/pull/10066) |
| implemented `DataModule`  `train_transforms`, `val_transforms`, `test_transforms`, `size`, `dims` | switch to `LightningDataModule` | [PR8851](https://github.com/Lightning-AI/pytorch-lightning/pull/8851) |

| If | Then | Ref |
| --- | --- | --- |
| used `DDP2Strategy` | switch to `DDPStrategy` | [PR14026](https://github.com/Lightning-AI/pytorch-lightning/pull/14026) |
| used  `Trainer.training_type_plugin` property | now use `Trainer.strategy` and update the references | [PR11141](https://github.com/Lightning-AI/pytorch-lightning/pull/11141) |
| used any  `TrainingTypePluginsn` | rename them to  `Strategy` | [PR11120](https://github.com/Lightning-AI/pytorch-lightning/pull/11120) |
| used `DistributedType` | rely on protected `_StrategyType` | [PR10505](https://github.com/Lightning-AI/pytorch-lightning/pull/10505) |
| used `DeviceType` | rely on protected  `_AcceleratorType` | [PR10503](https://github.com/Lightning-AI/pytorch-lightning/pull/10503) |
| used `pl.utiltiies.meta` functions | switch to built-in https://github.com/pytorch/torchdistx support | [PR13868](https://github.com/Lightning-AI/pytorch-lightning/pull/13868) |
| have implemented `Callback.on_configure_sharded_model` hook | move your implementation to `Callback.setup` | [PR14834](https://github.com/Lightning-AI/pytorch-lightning/pull/14834) |
| have implemented the `Callback.on_before_accelerator_backend_setup` hook | move your implementation to `Callback.setup` | [PR14834](https://github.com/Lightning-AI/pytorch-lightning/pull/14834) |
| have implemented the `Callback.on_batch_start` hook | move your implementation to `Callback.on_train_batch_start` | [PR14834](https://github.com/Lightning-AI/pytorch-lightning/pull/14834) |
| have implemented the `Callback.on_batch_end` hook | move your implementation to `Callback.on_train_batch_end` | [PR14834](https://github.com/Lightning-AI/pytorch-lightning/pull/14834) |
| have implemented the `Callback.on_epoch_start` hook | move your implementation  to `Callback.on_train_epoch_start` , to `Callback.on_validation_epoch_start` , to `Callback.on_test_epoch_start` | [PR14834](https://github.com/Lightning-AI/pytorch-lightning/pull/14834) |
| have implemented the `Callback.on_pretrain_routine_{start,end}` hook | move your implementation to `Callback.on_fit_start` | [PR14834](https://github.com/Lightning-AI/pytorch-lightning/pull/14834) |
| used `Callback.on_init_start` hook | use `Callback.on_train_start` instead | [PR10940](https://github.com/Lightning-AI/pytorch-lightning/pull/10940) |
| used `Callback.on_init_end`  hook | use `Callback.on_train_start` instead | [PR10940](https://github.com/Lightning-AI/pytorch-lightning/pull/10940) |
| used Trainer’s attribute `Trainer.num_processes` | it was replaced by  `Trainer.num_devices` | [PR12388](https://github.com/Lightning-AI/pytorch-lightning/pull/12388) |
| used Trainer’s attribute `Trainer.gpus` | it was replaced by  `Trainer.num_devices` | [PR12436](https://github.com/Lightning-AI/pytorch-lightning/pull/12436) |
| used Trainer’s attribute `Trainer.num_gpus` | use `Trainer.num_devices`  instead | [PR12384](https://github.com/Lightning-AI/pytorch-lightning/pull/12384) |
| used Trainer’s attribute `Trainer.ipus` | use  `Trainer.num_devices`  instead | [PR12386](https://github.com/Lightning-AI/pytorch-lightning/pull/12386) |
| used Trainer’s attribute `Trainer.tpu_cores` | use `Trainer.num_devices` instead | [PR12437](https://github.com/Lightning-AI/pytorch-lightning/pull/12437) |
| used  `Trainer.num_processes` attribute | switch to using `Trainer.num_devices` | [PR12388](https://github.com/Lightning-AI/pytorch-lightning/pull/12388) |
| used `LightningIPUModule` | it was removed | [PR14830](https://github.com/Lightning-AI/pytorch-lightning/pull/14830) |
| logged with `LightningLoggerBase.agg_and_log_metrics` | switch to `LightningLoggerBase.log_metrics` | [PR11832](https://github.com/Lightning-AI/pytorch-lightning/pull/11832) |
| used  `agg_key_funcs`  parameter from `LightningLoggerBase` | log metrics explicitly | [PR11871](https://github.com/Lightning-AI/pytorch-lightning/pull/11871) |
| used  `agg_default_func` parameters in `LightningLoggerBase` | log metrics explicitly | [PR11871](https://github.com/Lightning-AI/pytorch-lightning/pull/11871) |
| used  `Trainer.validated_ckpt_path` attribute | rely on generic read-only property `Trainer.ckpt_path` which is set when checkpoints are loaded via `Trainer.validate(ckpt_path=...)` | [PR11696](https://github.com/Lightning-AI/pytorch-lightning/pull/11696) |
| used  `Trainer.tested_ckpt_path` attribute | rely on generic read-only property `Trainer.ckpt_path` which is set when checkpoints are loaded via `Trainer.test(ckpt_path=...)` | [PR11696](https://github.com/Lightning-AI/pytorch-lightning/pull/11696) |
| used  `Trainer.predicted_ckpt_path` attribute | rely on generic read-only property `Trainer.ckpt_path`, which is set when checkpoints are loaded via `Trainer.predict(ckpt_path=...)` | [PR11696](https://github.com/Lightning-AI/pytorch-lightning/pull/11696) |
| rely on the returned dictionary from  `Callback.on_save_checkpoint` | call directly `Callback.state_dict` instead | [PR11887](https://github.com/Lightning-AI/pytorch-lightning/pull/11887) |

| If | Then | Ref |
| --- | --- | --- |
| imported `pl.callbacks.base` | import `pl.callbacks.callback` | [PR13031](https://github.com/Lightning-AI/pytorch-lightning/pull/13031) |
| imported `pl.loops.base` | import `pl.loops.loop` instead | [PR13043](https://github.com/Lightning-AI/pytorch-lightning/pull/13043) |
| imported `pl.utilities.cli` | import  `pl.cli` instead | [PR13767](https://github.com/Lightning-AI/pytorch-lightning/pull/13767) |
| imported profiler classes from `pl.profiler.*` | import `pl.profilers` instead | [PR12308](https://github.com/Lightning-AI/pytorch-lightning/pull/12308) |
| used `pl.accelerators.GPUAccelerator` | use `pl.accelerators.CUDAAccelerator` | [PR13636](https://github.com/Lightning-AI/pytorch-lightning/pull/13636) |
| used `LightningDeepSpeedModule` | use `strategy="deepspeed"` or `strategy=DeepSpeedStrategy(...)` | `DeepSpeedStrategy` |
| used the `with init_meta_context()` context manager from `import pl.utilities.meta` | switch to `deepspeed-zero-stage-3` | [deepspeed-zero-stage-3](/docs/pytorch/2.6.6/reference/advanced/model_parallel/deepspeed.md#deepspeed-zero-stage-3) |
| used the Lightning Hydra multi-run integration | removed support for it as it caused issues with processes hanging | [PR15689](https://github.com/Lightning-AI/pytorch-lightning/pull/15689) |
| used `pl.utilities.memory.get_gpu_memory_map` | use  `pl.accelerators.cuda.get_nvidia_gpu_stats` | [PR9921](https://github.com/Lightning-AI/pytorch-lightning/pull/9921) |

| If | Then | Ref |
| --- | --- | --- |
| used the `pl.lite` module | switch to `lightning_fabric` | [PR15953](https://github.com/Lightning-AI/pytorch-lightning/pull/15953) |
| used Trainer’s flag `strategy='dp'` | use DDP with `strategy='ddp'` or DeepSpeed instead | [PR16748](https://github.com/Lightning-AI/pytorch-lightning/pull/16748) |
| implemented `LightningModule.training_epoch_end` hooks | port your logic to  `LightningModule.on_train_epoch_end` hook | [PR16520](https://github.com/Lightning-AI/pytorch-lightning/pull/16520) |
| implemented `LightningModule.validation_epoch_end` hook | port your logic to  `LightningModule.on_validation_epoch_end` hook | [PR16520](https://github.com/Lightning-AI/pytorch-lightning/pull/16520) |
| implemented `LightningModule.test_epoch_end` hooks | port your logic to  `LightningModule.on_test_epoch_end` hook | [PR16520](https://github.com/Lightning-AI/pytorch-lightning/pull/16520) |
| used Trainer’s flag `multiple_trainloader_mode` | switch to  `CombinedLoader(..., mode=...)` and set mode directly now | [PR16800](https://github.com/Lightning-AI/pytorch-lightning/pull/16800) |
| used Trainer’s flag `move_metrics_to_cpu` | implement particular offload logic in your custom metric or turn it on in `torchmetrics` | [PR16358](https://github.com/Lightning-AI/pytorch-lightning/pull/16358) |
| used Trainer’s flag `track_grad_norm` | overwrite `on_before_optimizer_step` hook and pass the argument directly and `LightningModule.log_grad_norm()` hook | [PR16745](https://github.com/Lightning-AI/pytorch-lightning/pull/16745) [PR16745](https://github.com/Lightning-AI/pytorch-lightning/pull/16745) |
| used Trainer’s flag `replace_sampler_ddp` | use  `use_distributed_sampler`; the sampler gets created not only for the DDP strategies |  |
| relied on the `on_tpu` argument in `LightningModule.optimizer_step` hook | switch to manual optimization | [PR16537](https://github.com/Lightning-AI/pytorch-lightning/pull/16537) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| relied on the `using_lbfgs` argument in `LightningModule.optimizer_step` hook | switch to manual optimization | [PR16538](https://github.com/Lightning-AI/pytorch-lightning/pull/16538) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| were using `nvidia/apex` in any form | switch to PyTorch native mixed precision `torch.amp` instead | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| used Trainer’s flag `using_native_amp` | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| used Trainer’s flag `amp_backend` | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| used Trainer’s flag `amp_level` | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| used Trainer’s attribute `using_native_amp` | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| used Trainer’s attribute `amp_backend` | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| used Trainer’s attribute `amp_level` | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) [Precision](/docs/pytorch/2.6.6/reference/common/precision.md) |
| use the `FairScale` integration | consider using PyTorch's native FSDP implementation or outsourced implementation into own project | [lightning-Fairscale](https://github.com/Lightning-Sandbox/lightning-Fairscale) |
| used `pl.overrides.fairscale.LightningShardedDataParallel` | use native FSDP instead | [PR16400](https://github.com/Lightning-AI/pytorch-lightning/pull/16400) [FSDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `pl.plugins.precision.fully_sharded_native_amp.FullyShardedNativeMixedPrecisionPlugin` | use native FSDP instead | [PR16400](https://github.com/Lightning-AI/pytorch-lightning/pull/16400) [FSDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `pl.plugins.precision.sharded_native_amp.ShardedNativeMixedPrecisionPlugin` | use native FSDP instead | [PR16400](https://github.com/Lightning-AI/pytorch-lightning/pull/16400) [FSDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `pl.strategies.fully_sharded.DDPFullyShardedStrategy` | use native FSDP instead | [PR16400](https://github.com/Lightning-AI/pytorch-lightning/pull/16400) [FSDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `pl.strategies.sharded.DDPShardedStrategy` | use native FSDP instead | [PR16400](https://github.com/Lightning-AI/pytorch-lightning/pull/16400) [FSDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `pl.strategies.sharded_spawn.DDPSpawnShardedStrategy` | use native FSDP instead | [PR16400](https://github.com/Lightning-AI/pytorch-lightning/pull/16400) [FSDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `save_config_overwrite` parameters in `LightningCLI` | pass this option and via dictionary of `save_config_kwargs` parameter | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) |
| used `save_config_multifile` parameters in `LightningCLI` | pass this option and via dictionary of `save_config_kwargs` parameter | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) |
| have customized loops `Loop.replace()` | implement your training loop with Fabric. | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| have customized loops `Loop.run()` | implement your training loop with Fabric. | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| have customized loops `Loop.connect()` | implement your training loop with Fabric. | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| used the Trainer’s `trainer.fit_loop` property | implement your training loop with Fabric | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| used the Trainer’s `trainer.validate_loop` property | implement your training loop with Fabric | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| used the Trainer’s `trainer.test_loop` property | implement your training loop with Fabric | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| used the Trainer’s `trainer.predict_loop` property | implement your training loop with Fabric | [PR14998](https://github.com/Lightning-AI/pytorch-lightning/pull/14998) [Fabric](/docs/fabric) |
| used the `Trainer.loop` and fetching classes | being marked as protected |  |
| used `opt_idx` argument in `BaseFinetuning.finetune_function` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) |
| used `opt_idx` argument in `Callback.on_before_optimizer_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` as an optional argument in `LightningModule.training_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `LightningModule.on_before_optimizer_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `LightningModule.configure_gradient_clipping` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `LightningModule.optimizer_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `LightningModule.optimizer_zero_grad` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `LightningModule.lr_scheduler_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used declaring optimizer frequencies in the dictionary returned from `LightningModule.configure_optimizers` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer` argument in `LightningModule.backward` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `LightningModule.backward` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `PrecisionPlugin.optimizer_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `PrecisionPlugin.,backward` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `PrecisionPlugin.optimizer_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `Strategy.backward` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `optimizer_idx` argument in `Strategy.optimizer_step` | use manual optimization | [PR16539](https://github.com/Lightning-AI/pytorch-lightning/pull/16539) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used Trainer’s `Trainer.optimizer_frequencies` attribute | use manual optimization | [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `PL_INTER_BATCH_PARALLELISM` environment flag |  | [PR16355](https://github.com/Lightning-AI/pytorch-lightning/pull/16355) |
| used training integration with Horovod | install standalone package/project | [lightning-Horovod](https://github.com/Lightning-AI/lightning-Horovod) |
| used training integration with ColossalAI | install standalone package/project | [lightning-ColossalAI](/docs/pytorch/2.1.0/integrations/strategies/colossalai) |
| used `QuantizationAwareTraining` callback | use Torch’s Quantization directly | [PR16750](https://github.com/Lightning-AI/pytorch-lightning/pull/16750) |
| had any logic except reducing the DP outputs in  `LightningModule.training_step_end` hook | port it to `LightningModule.on_train_batch_end` hook | [PR16791](https://github.com/Lightning-AI/pytorch-lightning/pull/16791) |
| had any logic except reducing the DP outputs in  `LightningModule.validation_step_end` hook | port it to `LightningModule.on_validation_batch_end` hook | [PR16791](https://github.com/Lightning-AI/pytorch-lightning/pull/16791) |
| had any logic except reducing the DP outputs in  `LightningModule.test_step_end` hook | port it to `LightningModule.on_test_batch_end` hook | [PR16791](https://github.com/Lightning-AI/pytorch-lightning/pull/16791) |
| used `pl.strategies.DDPSpawnStrategy` | switch to general  `DDPStrategy(start_method='spawn')` with proper starting method | [PR16809](https://github.com/Lightning-AI/pytorch-lightning/pull/16809) |
| used the automatic addition of a moving average of the `training_step` loss in the progress bar | use `self.log("loss", ..., prog_bar=True)` instead. | [PR16192](https://github.com/Lightning-AI/pytorch-lightning/pull/16192) |
| rely on the `outputs` argument from the `on_predict_epoch_end` hook | access them via `trainer.predict_loop.predictions` | [PR16655](https://github.com/Lightning-AI/pytorch-lightning/pull/16655) |
| need to pass a dictionary to `self.log()` | pass them independently. | [PR16389](https://github.com/Lightning-AI/pytorch-lightning/pull/16389) |

## Developer {#developer}

| If | Then | Ref |
| --- | --- | --- |
| called `CheckpointConnector.hpc_load()` | just call `CheckpointConnector.restore()` | [PR7652](https://github.com/Lightning-AI/pytorch-lightning/pull/7652) |
| used `TrainerModelHooksMixin` | now rely on the corresponding utility functions in `pytorch_lightning.utilities.signature_utils` | [PR7422](https://github.com/Lightning-AI/pytorch-lightning/pull/7422) |
| assigned the `Trainer.train_loop` property | now assign the equivalent `Trainer.fit_loop` property | [PR8025](https://github.com/Lightning-AI/pytorch-lightning/pull/8025) |
| accessed `LightningModule.loaded_optimizer_states_dict` | the property has been removed | [PR8229](https://github.com/Lightning-AI/pytorch-lightning/pull/8229) |

| If | Then | Ref |
| --- | --- | --- |
| called `LightningLoggerBase.close` | switch to `LightningLoggerBase.finalize`. | [PR9422](https://github.com/Lightning-AI/pytorch-lightning/pull/9422) |
| called `LoggerCollection.close` | switch to `LoggerCollection.finalize`. | [PR9422](https://github.com/Lightning-AI/pytorch-lightning/pull/9422) |
| used `AcceleratorConnector.is_slurm_managing_tasks` attribute | it is set not as protected and discouraged from direct use | [PR10101](https://github.com/Lightning-AI/pytorch-lightning/pull/10101) |
| used `AcceleratorConnector.configure_slurm_ddp` attributes | it is set not as protected and discouraged from direct use | [PR10101](https://github.com/Lightning-AI/pytorch-lightning/pull/10101) |
| used `ClusterEnvironment.creates_children()` method | change it to `ClusterEnvironment.creates_processes_externally` which is property now. | [PR10106](https://github.com/Lightning-AI/pytorch-lightning/pull/10106) |
| called `PrecisionPlugin.master_params()` | update it  `PrecisionPlugin.main_params()` | [PR10105](https://github.com/Lightning-AI/pytorch-lightning/pull/10105) |

| If | Then | Ref |
| --- | --- | --- |
| Removed the legacy `Trainer.get_deprecated_arg_names()` |  | [PR14415](https://github.com/Lightning-AI/pytorch-lightning/pull/14415) |
| used the generic method `Trainer.run_stage` | switch to a specific one depending on your purpose `Trainer.{fit,validate,test,predict}` . | [PR11000](https://github.com/Lightning-AI/pytorch-lightning/pull/11000) |
| used `rank_zero_only` from `pl.utilities.distributed` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `rank_zero_debug` from `pl.utilities.distributed` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `rank_zero_info` from `pl.utilities.distributed` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `rank_zero_warn` from `pl.utilities.warnings` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `rank_zero_deprecation` from `pl.utilities.warnings` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `LightningDeprecationWarning` from `pl.utilities.warnings` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `LightningDeprecationWarning` from `pl.utilities.warnings` | import it from `pl.utilities.rank_zero` | [PR11747](https://github.com/Lightning-AI/pytorch-lightning/pull/11747) |
| used `Trainer.data_parallel_device_ids` attribute | switch it to `Trainer.device_ids` | [PR12072](https://github.com/Lightning-AI/pytorch-lightning/pull/12072) |
| derived it from `TrainerCallbackHookMixin` | use Trainer base class | [PR14401](https://github.com/Lightning-AI/pytorch-lightning/pull/14401) |
| used base class `pytorch_lightning.profiler.BaseProfilerto` | switch to use `pytorch_lightning.profiler.Profiler` instead | [PR12150](https://github.com/Lightning-AI/pytorch-lightning/pull/12150) |
| set distributed backend via the environment variable `PL_TORCH_DISTRIBUTED_BACKEND` | use `process_group_backend` in the strategy constructor | [PR11745](https://github.com/Lightning-AI/pytorch-lightning/pull/11745) |
| used `PrecisionPlugin.on_load_checkpoint` hooks | switch to  `PrecisionPlugin.load_state_dict` | [PR11978](https://github.com/Lightning-AI/pytorch-lightning/pull/11978) |
| used `PrecisionPlugin.on_save_checkpoint` hooks | switch to  `PrecisionPlugin.load_state_dict` | [PR11978](https://github.com/Lightning-AI/pytorch-lightning/pull/11978) |
| used `Trainer.root_gpu` attribute | use `Trainer.strategy.root_device.index` when GPU is used | [PR12262](https://github.com/Lightning-AI/pytorch-lightning/pull/12262) |
| used `Trainer.use_amp` attribute | rely on Torch native AMP | [PR12312](https://github.com/Lightning-AI/pytorch-lightning/pull/12312) |
| used `LightningModule.use_amp` attribute | rely on Torch native AMP | [PR12315](https://github.com/Lightning-AI/pytorch-lightning/pull/12315) |
| used Trainer’s attribute `Trainer.verbose_evaluate` | rely on loop constructor  `EvaluationLoop(verbose=...)` | [PR10931](https://github.com/Lightning-AI/pytorch-lightning/pull/10931) |
| used Trainer’s attribute `Trainer.should_rank_save_checkpoint` | it was removed | [PR11068](https://github.com/Lightning-AI/pytorch-lightning/pull/11068) |
| derived from `TrainerOptimizersMixin` | rely on `core/optimizer.py` | [PR11155](https://github.com/Lightning-AI/pytorch-lightning/pull/11155) |
| derived from `TrainerDataLoadingMixin` | rely on methods from `Trainer` and `DataConnector` | [PR11282](https://github.com/Lightning-AI/pytorch-lightning/pull/11282) |
| used Trainer’s attribute `Trainer.lightning_optimizers` | switch to the `Strategy` and its attributes. | [PR11444](https://github.com/Lightning-AI/pytorch-lightning/pull/11444) |
| used `Trainer.call_hook` | it was set as a protected method `Trainer._call_callback_hooks`, `Trainer._call_lightning_module_hook`, `Trainer._call_ttp_hook`, `Trainer._call_accelerator_hook` and shall not be used. | [PR10979](https://github.com/Lightning-AI/pytorch-lightning/pull/10979) |
| used Profiler’s attribute  `SimpleProfiler.profile_iterable` | it was removed | [PR12102](https://github.com/Lightning-AI/pytorch-lightning/pull/12102) |
| used Profiler’s attribute  `AdvancedProfiler.profile_iterable` | it was removed | [PR12102](https://github.com/Lightning-AI/pytorch-lightning/pull/12102) |
| used the  `device_stats_monitor.prefix_metric_keys` |  | [PR11254](https://github.com/Lightning-AI/pytorch-lightning/pull/11254) |
| used `on_train_batch_end(outputs, ...)` with 2d list with sizes (n_optimizers, tbptt_steps) | chang it to (tbptt_steps, n_optimizers). You can update your code by adding the following parameter to your hook signature: `on_train_batch_end(outputs, ..., new_format=True)`. | [PR12182](https://github.com/Lightning-AI/pytorch-lightning/pull/12182) |
| used `training_epoch_end(outputs)` with a 3d list with sizes (n_optimizers, n_batches, tbptt_steps) | change it to (n_batches, tbptt_steps, n_optimizers). You can update your code by adding the following parameter to your hook signature: `training_epoch_end(outputs, new_format=True)`. | [PR12182](https://github.com/Lightning-AI/pytorch-lightning/pull/12182) |

| If | Then | Ref |
| --- | --- | --- |
| derived from `pytorch_lightning.loggers.base.LightningLoggerBase` | derive from `pytorch_lightning.loggers.logger.Logger` | [PR12014](https://github.com/Lightning-AI/pytorch-lightning/pull/12014) |
| derived from `pytorch_lightning.profiler.base.BaseProfiler` | derive from `pytorch_lightning.profilers.profiler.Profiler` | [PR12150](https://github.com/Lightning-AI/pytorch-lightning/pull/12150) |
| derived from `pytorch_lightning.profiler.base.AbstractProfiler` | derive from `pytorch_lightning.profilers.profiler.Profiler` | [PR12106](https://github.com/Lightning-AI/pytorch-lightning/pull/12106) |

| If | Then | Ref |
| --- | --- | --- |
| passed the `pl_module` argument to distributed module wrappers | passed the (required) `forward_module` argument | [PR16386](https://github.com/Lightning-AI/pytorch-lightning/pull/16386) |
| used `DataParallel` and the `LightningParallelModule` wrapper | use DDP or DeepSpeed instead | [PR16748](https://github.com/Lightning-AI/pytorch-lightning/pull/16748) [DDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used `pl_module` argument from the distributed module wrappers | use DDP or DeepSpeed instead | [PR16386](https://github.com/Lightning-AI/pytorch-lightning/pull/16386) [DDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| called `pl.overrides.base.unwrap_lightning_module` function | use DDP or DeepSpeed instead | [PR16386](https://github.com/Lightning-AI/pytorch-lightning/pull/16386) [DDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used or derived from `pl.overrides.distributed.LightningDistributedModule` class | use DDP instead | [PR16386](https://github.com/Lightning-AI/pytorch-lightning/pull/16386) [DDP](/docs/pytorch/2.6.6/reference/accelerators/gpu_expert.md) |
| used the `pl.plugins.ApexMixedPrecisionPlugin` plugin | use PyTorch native mixed precision | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) |
| used the `pl.plugins.NativeMixedPrecisionPlugin` plugin | switch to the `pl.plugins.MixedPrecisionPlugin` plugin | [PR16039](https://github.com/Lightning-AI/pytorch-lightning/pull/16039) |
| used the `fit_loop.min_steps` setters | implement your training loop with Fabric | [PR16803](https://github.com/Lightning-AI/pytorch-lightning/pull/16803) |
| used the `fit_loop.max_steps` setters | implement your training loop with Fabric | [PR16803](https://github.com/Lightning-AI/pytorch-lightning/pull/16803) |
| used the `data_parallel` attribute in `Trainer` | check the same using `isinstance(trainer.strategy, ParallelStrategy)` | [PR16703](https://github.com/Lightning-AI/pytorch-lightning/pull/16703) |
| used any function from `pl.utilities.xla_device` | switch to `pl.accelerators.XLAAccelerator.is_available()` | [PR14514](https://github.com/Lightning-AI/pytorch-lightning/pull/14514) [PR14550](https://github.com/Lightning-AI/pytorch-lightning/pull/14550) |
| imported functions from  `pl.utilities.device_parser.*` | import them from `lightning_fabric.utilities.device_parser.*` | [PR14492](https://github.com/Lightning-AI/pytorch-lightning/pull/14492) [PR14753](https://github.com/Lightning-AI/pytorch-lightning/pull/14753) |
| imported functions from `pl.utilities.cloud_io.*` | import them from `lightning_fabric.utilities.cloud_io.*` | [PR14515](https://github.com/Lightning-AI/pytorch-lightning/pull/14515) |
| imported functions from `pl.utilities.apply_func.*` | import them from `lightning_utilities.core.apply_func.*` | [PR14516](https://github.com/Lightning-AI/pytorch-lightning/pull/14516) [PR14537](https://github.com/Lightning-AI/pytorch-lightning/pull/14537) |
| used any code from `pl.core.mixins` | use the base classes | [PR16424](https://github.com/Lightning-AI/pytorch-lightning/pull/16424) |
| used any code from `pl.utilities.distributed` | rely on Pytorch's native functions | [PR16390](https://github.com/Lightning-AI/pytorch-lightning/pull/16390) |
| used any code from `pl.utilities.data` | it was removed | [PR16440](https://github.com/Lightning-AI/pytorch-lightning/pull/16440) |
| used any code from `pl.utilities.optimizer` | it was removed | [PR16439](https://github.com/Lightning-AI/pytorch-lightning/pull/16439) |
| used any code from `pl.utilities.seed` | it was removed | [PR16422](https://github.com/Lightning-AI/pytorch-lightning/pull/16422) |
| were using truncated backpropagation through time (TBPTT) with `LightningModule.truncated_bptt_steps` | use manual optimization | [PR16172](https://github.com/Lightning-AI/pytorch-lightning/pull/16172) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| were using truncated backpropagation through time (TBPTT) with `LightningModule.tbptt_split_batch` | use manual optimization | [PR16172](https://github.com/Lightning-AI/pytorch-lightning/pull/16172) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| were using truncated backpropagation through time (TBPTT) and passing `hidden`  to `LightningModule.training_step` | use manual optimization | [PR16172](https://github.com/Lightning-AI/pytorch-lightning/pull/16172) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `pl.utilities.finite_checks.print_nan_gradients` function | it was removed |  |
| used `pl.utilities.finite_checks.detect_nan_parameters` function | it was removed |  |
| used `pl.utilities.parsing.flatten_dict` function | it was removed |  |
| used `pl.utilities.metrics.metrics_to_scalars` function | it was removed |  |
| used `pl.utilities.memory.get_model_size_mb` function | it was removed |  |
| used `pl.strategies.utils.on_colab_kaggle` function | it was removed | [PR16437](https://github.com/Lightning-AI/pytorch-lightning/pull/16437) |
| used `LightningDataModule.add_argparse_args()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `LightningDataModule.parse_argparser()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `LightningDataModule.from_argparse_args()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `LightningDataModule.get_init_arguments_and_types()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `Trainer.default_attributes()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `Trainer.from_argparse_args()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `Trainer.parse_argparser()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `Trainer.match_env_arguments()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `Trainer.add_argparse_args()` method | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.argparse.from_argparse_args()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.argparse.parse_argparser()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.argparseparse_env_variables()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `get_init_arguments_and_types()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.argparse.add_argparse_args()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.parsing.str_to_bool()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.parsing.str_to_bool_or_int()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| used `pl.utilities.parsing.str_to_bool_or_str()` function | switch to using `LightningCLI` | [PR16708](https://github.com/Lightning-AI/pytorch-lightning/pull/16708) |
| derived from `pl.utilities.distributed.AllGatherGrad` class | switch to PyTorch native equivalent | [PR15364](https://github.com/Lightning-AI/pytorch-lightning/pull/15364) |
| used `PL_RECONCILE_PROCESS=1` env. variable | customize your logger | [PR16204](https://github.com/Lightning-AI/pytorch-lightning/pull/16204) |
| if you derived from mixin’s method `pl.core.saving.ModelIO.load_from_checkpoint` | rely on `pl.core.module.LightningModule` | [PR16999](https://github.com/Lightning-AI/pytorch-lightning/pull/16999) |
| used  `Accelerator.setup_environment` method | switch to `Accelerator.setup_device` | [PR16436](https://github.com/Lightning-AI/pytorch-lightning/pull/16436) |
| used `PL_FAULT_TOLERANT_TRAINING` env. variable | implement own logic with Fabric | [PR16516](https://github.com/Lightning-AI/pytorch-lightning/pull/16516) [PR16533](https://github.com/Lightning-AI/pytorch-lightning/pull/16533) |
| used or derived from public `pl.overrides.distributed.IndexBatchSamplerWrapper` class | it is set as protected | [PR16826](https://github.com/Lightning-AI/pytorch-lightning/pull/16826) |
| used the `DataLoaderLoop` class | use manual optimization | [PR16726](https://github.com/Lightning-AI/pytorch-lightning/pull/16726) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used the `EvaluationEpochLoop` class | use manual optimization | [PR16726](https://github.com/Lightning-AI/pytorch-lightning/pull/16726) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used the `PredictionEpochLoop` class | use manual optimization | [PR16726](https://github.com/Lightning-AI/pytorch-lightning/pull/16726) [Manual Optimization](/docs/pytorch/2.6.6/reference/model/manual_optimization.md) |
| used `trainer.reset_*_dataloader()` methods | use  `Loop.setup_data()` for the top-level loops | [PR16726](https://github.com/Lightning-AI/pytorch-lightning/pull/16726) |
| used `LightningModule.precision` attribute | rely on Trainer precision attribute | [PR16203](https://github.com/Lightning-AI/pytorch-lightning/pull/16203) |
| used  `Trainer.model` setter | you shall pass the `model` in fit/test/predict method | [PR16462](https://github.com/Lightning-AI/pytorch-lightning/pull/16462) |
| relied on `pl.utilities.supporters.CombinedLoaderIterator` class | pass dataloders directly | [PR16714](https://github.com/Lightning-AI/pytorch-lightning/pull/16714) |
| relied on `pl.utilities.supporters.CombinedLoaderIterator` class | pass dataloders directly | [PR16714](https://github.com/Lightning-AI/pytorch-lightning/pull/16714) |
| used `pl.callbacks.progress.base.ProgressBarBase` | rename to `pl.callbacks.progress.ProgressBar` | [PR17058](https://github.com/Lightning-AI/pytorch-lightning/pull/17058) |
| accessed `ProgressBarBase.train_batch_idx` property | rely on Trainer internal loops’ properties | [PR16760](https://github.com/Lightning-AI/pytorch-lightning/pull/16760) |
| accessed `ProgressBarBase.val_batch_idx` property | rely on Trainer internal loops’ properties | [PR16760](https://github.com/Lightning-AI/pytorch-lightning/pull/16760) |
| accessed `ProgressBarBase.test_batch_idx` property | rely on Trainer internal loops’ properties | [PR16760](https://github.com/Lightning-AI/pytorch-lightning/pull/16760) |
| accessed `ProgressBarBase.predict_batch_idx` property | rely on Trainer internal loops’ properties | [PR16760](https://github.com/Lightning-AI/pytorch-lightning/pull/16760) |
| used `Trainer.prediction_writer_callbacks` property | rely on precision plugin | [PR16759](https://github.com/Lightning-AI/pytorch-lightning/pull/16759) |
| used `PrecisionPlugin.dispatch` | it was removed | [PR16618](https://github.com/Lightning-AI/pytorch-lightning/pull/16618) |
| used `Strategy.dispatch` | it was removed | [PR16618](https://github.com/Lightning-AI/pytorch-lightning/pull/16618) |

---

**Navigation**

- Parent: [2.0 Upgrade Guide](/docs/pytorch/2.6.6/home/migration_guide.md)