Hardware specific parts of Accelerator Refactoring

CHANGELOG.md

@@ -110,7 +110,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Refactored Accelerators and Plugins 
     * Added base classes for plugins ([#5715](https://github.com/PyTorchLightning/pytorch-lightning/pull/5715))
     * Added parallel plugins for DP, DDP, DDPSpawn, DDP2 and Horovod ([#5714](https://github.com/PyTorchLightning/pytorch-lightning/pull/5714))
-
+    * Added new Accelerators for CPU, GPU and TPU ([#5719](https://github.com/PyTorchLightning/pytorch-lightning/pull/5719))
+    * Added Plugins for TPU training ([#5719](https://github.com/PyTorchLightning/pytorch-lightning/pull/5719))
 
 ### Deprecated
 

docs/source/common/trainer.rst

@@ -1121,7 +1121,7 @@ To define your own behavior, subclass the relevant class and pass it in. Here's
 
 .. code-block:: python
 
-    from pytorch_lightning.cluster_environments import cluster_environment
+    from pytorch_lightning.plugins.environments import cluster_environment
 
     class MyCluster(ClusterEnvironment):
 

docs/source/extensions/accelerators.rst

@@ -70,7 +70,7 @@ First, implement your own ClusterEnvironment. Here is the torch elastic implemen
     import os
     from pytorch_lightning import _logger as log
     from pytorch_lightning.utilities import rank_zero_warn
-    from pytorch_lightning.cluster_environments.cluster_environment import ClusterEnvironment
+    from pytorch_lightning.plugins.environments.cluster_environment import ClusterEnvironment
 
     class TorchElasticEnvironment(ClusterEnvironment):
 

pytorch_lightning/accelerators/accelerator.py

@@ -347,8 +347,7 @@ def amp_backend(self) -> Optional[LightningEnum]:
             return AMPType.APEX
         elif isinstance(self.precision_plugin, NativeMixedPrecisionPlugin):
             return AMPType.NATIVE
-        else:
-            return None
+        return None
 
     @property
     def precision(self) -> int: