diff --git a/pytorch_lightning/plugins/training_type/parallel.py b/pytorch_lightning/plugins/training_type/parallel.py
index e6406ea444..4f4b2c5b8e 100644
--- a/pytorch_lightning/plugins/training_type/parallel.py
+++ b/pytorch_lightning/plugins/training_type/parallel.py
@@ -44,7 +44,7 @@ class ParallelPlugin(TrainingTypePlugin, ABC):
     @property
     @abstractmethod
     def root_device(self) -> torch.device:
-        raise NotImplementedError
+        """Return the root device."""
 
     @property
     def on_gpu(self) -> bool:
diff --git a/pytorch_lightning/plugins/training_type/training_type_plugin.py b/pytorch_lightning/plugins/training_type/training_type_plugin.py
index 585c8b6405..a1e55631c2 100644
--- a/pytorch_lightning/plugins/training_type/training_type_plugin.py
+++ b/pytorch_lightning/plugins/training_type/training_type_plugin.py
@@ -86,19 +86,16 @@ class TrainingTypePlugin(ABC):
     @abstractmethod
     def on_gpu(self) -> bool:
         """Returns whether the current process is done on GPU."""
-        raise NotImplementedError
 
     @property
     @abstractmethod
     def on_tpu(self) -> bool:
         """Returns whether the current process is done on TPU."""
-        raise NotImplementedError
 
     @property
     @abstractmethod
     def root_device(self) -> torch.device:
         """Returns the root device."""
-        raise NotImplementedError
 
     @abstractmethod
     def model_to_device(self) -> None:
@@ -321,7 +318,6 @@ class TrainingTypePlugin(ABC):
 
         It is the right place to release memory and free other resources.
         """
-        raise NotImplementedError
 
     @classmethod
     def register_plugins(cls, plugin_registry) -> None: