Deprecate Trainer.should_rank_save_checkpoint property (#11068)

Co-authored-by: Carlos Mocholí <[email protected]>

Author: jjenniferdai

CHANGELOG.md

@@ -147,6 +147,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Deprecated `ModelIO.on_hpc_{save/load}` in favor of `CheckpointHooks.on_{save/load}_checkpoint` ([#10911](https://github.com/PyTorchLightning/pytorch-lightning/pull/10911))
 
 
+- Deprecated `Trainer.should_rank_save_checkpoint` Trainer property ([#11068](https://github.com/PyTorchLightning/pytorch-lightning/pull/11068))
+
+
 ### Removed
 
 - Removed deprecated parameter `method` in `pytorch_lightning.utilities.model_helpers.is_overridden` ([#10507](https://github.com/PyTorchLightning/pytorch-lightning/pull/10507))
@@ -611,7 +614,6 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Removed `TrainerProperties` mixin and moved property definitions directly into `Trainer` ([#9495](https://github.com/PyTorchLightning/pytorch-lightning/pull/9495))
 - Removed a redundant warning with `ModelCheckpoint(monitor=None)` callback ([#9875](https://github.com/PyTorchLightning/pytorch-lightning/pull/9875))
 - Remove `epoch` from `trainer.logged_metrics` ([#9904](https://github.com/PyTorchLightning/pytorch-lightning/pull/9904))
-- Removed `should_rank_save_checkpoint` property from Trainer ([#9433](https://github.com/PyTorchLightning/pytorch-lightning/pull/9433))
 - Remove deprecated `distributed_backend` from `Trainer` ([#10017](https://github.com/PyTorchLightning/pytorch-lightning/pull/10017))
 - Removed `process_idx` from the `{DDPSpawnPlugin,TPUSpawnPlugin}.new_process` methods ([#10022](https://github.com/PyTorchLightning/pytorch-lightning/pull/10022))
 - Removed automatic patching of `{train,val,test,predict}_dataloader()` on the `LightningModule` ([#9764](https://github.com/PyTorchLightning/pytorch-lightning/pull/9764))

pytorch_lightning/trainer/trainer.py

@@ -1699,6 +1699,13 @@ def world_size(self) -> int:
         # some training types define a world size
         return getattr(self.training_type_plugin, "world_size", 1)
 
+    @property
+    def should_rank_save_checkpoint(self) -> bool:
+        rank_zero_deprecation(
+            "`Trainer.should_rank_save_checkpoint` is deprecated in v1.6 and will be removed in 1.8.", stacklevel=5
+        )
+        return self.training_type_plugin.should_rank_save_checkpoint
+
     @property
     def _distrib_type(self) -> _StrategyType:
         return self._accelerator_connector._distrib_type

tests/deprecated_api/test_remove_1-8.py

@@ -106,3 +106,11 @@ def on_hpc_load(self):
         match=r"Method `LightningModule.on_hpc_load` is deprecated in v1.6 and will be removed in v1.8."
     ):
         trainer.fit(load_model)
+
+
+def test_v1_8_0_deprecated_trainer_should_rank_save_checkpoint(tmpdir):
+    trainer = Trainer()
+    with pytest.deprecated_call(
+        match=r"`Trainer.should_rank_save_checkpoint` is deprecated in v1.6 and will be removed in 1.8."
+    ):
+        _ = trainer.should_rank_save_checkpoint