clean up unused attributes in LightningModule (#8259)

pytorch_lightning/core/lightning.py

@@ -91,8 +91,6 @@ class LightningModule(
         # torch/nn/modules/module.py#L227)
         torch._C._log_api_usage_once(f"lightning.module.{self.__class__.__name__}")
 
-        self._loaded_optimizer_states_dict = {}
-
         # pointer to the trainer object
         self.trainer = None
 
@@ -109,13 +107,15 @@ class LightningModule(
         self._example_input_array = None
         self._datamodule = None
         self._current_fx_name: Optional[str] = None
-        self._running_manual_backward: bool = False
         self._current_dataloader_idx: Optional[int] = None
         self._automatic_optimization: bool = True
         self._truncated_bptt_steps: int = 0
         self._param_requires_grad_state = dict()
         self._metric_attributes: Optional[Dict[int, str]] = None
 
+        # deprecated, will be removed in 1.6
+        self._loaded_optimizer_states_dict = {}
+
     def optimizers(self, use_pl_optimizer: bool = True) -> Union[Optimizer, List[Optimizer], List[LightningOptimizer]]:
         """
         Returns the optimizer(s) that are being used during training. Useful for manual optimization.
@@ -1450,9 +1450,7 @@ class LightningModule(
         self._verify_is_manual_optimization('manual_backward')
 
         # backward
-        self._running_manual_backward = True
         self.trainer.fit_loop.epoch_loop.batch_loop.backward(loss, optimizer=None, opt_idx=None, *args, **kwargs)
-        self._running_manual_backward = False
 
     def backward(self, loss: Tensor, optimizer: Optimizer, optimizer_idx: int, *args, **kwargs) -> None:
         """
@@ -1470,8 +1468,7 @@ class LightningModule(
             def backward(self, loss, optimizer, optimizer_idx):
                 loss.backward()
         """
-        if self.automatic_optimization or self._running_manual_backward:
-            loss.backward(*args, **kwargs)
+        loss.backward(*args, **kwargs)
 
     def toggle_optimizer(self, optimizer: Optimizer, optimizer_idx: int):
         """

pytorch_lightning/trainer/connectors/model_connector.py

@@ -24,7 +24,5 @@ class ModelConnector:
 
         for m in [model, ref_model]:
             m.trainer = proxy(self.trainer)
-            m._device_type = str(self.trainer._device_type)
-            m._distrib_type = str(self.trainer._distrib_type)
             m.use_amp = self.trainer.amp_backend is not None
             m.precision = self.trainer.precision