Remove requirement to call `Fabric.launch()` with DP strategy (#17931)

src/lightning/fabric/CHANGELOG.md

@@ -142,6 +142,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Fixed check for FSDP's flat parameters in all parameter groups ([#17914](https://github.com/Lightning-AI/lightning/pull/17914))
 
 
+- Removed the need to call `.launch()` when using the DP-strategy (`strategy="dp"`) ([#17931](https://github.com/Lightning-AI/lightning/pull/17931))
+
+
 - Fixed automatic step tracking in Fabric's CSVLogger ([#17942](https://github.com/Lightning-AI/lightning/pull/17942))
 
 

src/lightning/fabric/fabric.py

@@ -33,7 +33,14 @@ from lightning.fabric.utilities.imports import _TORCH_GREATER_EQUAL_2_0
 from lightning.fabric.plugins import Precision  # avoid circular imports: # isort: split
 from lightning.fabric.accelerators.accelerator import Accelerator
 from lightning.fabric.connector import _Connector, _is_using_cli, _PLUGIN_INPUT, _PRECISION_INPUT
-from lightning.fabric.strategies import DeepSpeedStrategy, FSDPStrategy, SingleDeviceStrategy, Strategy, XLAStrategy
+from lightning.fabric.strategies import (
+    DataParallelStrategy,
+    DeepSpeedStrategy,
+    FSDPStrategy,
+    SingleDeviceStrategy,
+    Strategy,
+    XLAStrategy,
+)
 from lightning.fabric.strategies.launchers import _MultiProcessingLauncher, _XLALauncher
 from lightning.fabric.strategies.strategy import _Sharded, TBroadcast
 from lightning.fabric.utilities import move_data_to_device
@@ -919,7 +926,7 @@ class Fabric:
         setattr(self, "run", partial(self._wrap_and_launch, self.run))
 
     def _validate_launched(self) -> None:
-        if not self._launched and not isinstance(self._strategy, SingleDeviceStrategy):
+        if not self._launched and not isinstance(self._strategy, (SingleDeviceStrategy, DataParallelStrategy)):
             raise RuntimeError(
                 "To use Fabric with more than one device, you must call `.launch()` or use the CLI:"
                 " `lightning run model --help`."

tests/tests_fabric/test_fabric.py

@@ -27,6 +27,7 @@ from torch.utils.data import DataLoader, DistributedSampler, RandomSampler, Samp
 from lightning.fabric.fabric import Fabric
 from lightning.fabric.plugins import Precision
 from lightning.fabric.strategies import (
+    DataParallelStrategy,
     DDPStrategy,
     DeepSpeedStrategy,
     ParallelStrategy,
@@ -1134,6 +1135,8 @@ def test_verify_launch_called():
     assert not fabric._launched
     fabric._strategy = Mock(spec=SingleDeviceStrategy)
     fabric._validate_launched()
+    fabric._strategy = Mock(spec=DataParallelStrategy)
+    fabric._validate_launched()
     fabric._strategy = Mock(spec=DDPStrategy)
     with pytest.raises(RuntimeError, match=r"you must call `.launch\(\)`"):
         fabric._validate_launched()