ref: reduced all simplified_forward (#3126)

* simplified training_forward

* simplified training_forward

* simplified training_forward

pytorch_lightning/trainer/training_loop.py

@@ -1162,16 +1162,7 @@ class TrainerTrainLoopMixin(ABC):
             model.cpu()
             torch.cuda.empty_cache()
 
-    def training_forward(self, batch, batch_idx, opt_idx, hiddens):
-        """
-        Handle forward for each training case (distributed, single gpu, etc...)
-        :param batch:
-        :param batch_idx:
-        :return:
-        """
-        # ---------------
-        # FORWARD
-        # ---------------
+    def build_train_args(self, batch, batch_idx, opt_idx, hiddens):
         # enable not needing to add opt_idx to training_step
         args = [batch, batch_idx]
 
@@ -1189,24 +1180,21 @@ class TrainerTrainLoopMixin(ABC):
         if self.truncated_bptt_steps is not None:
             args.append(hiddens)
 
+        return args
+
+    def training_forward(self, batch, batch_idx, opt_idx, hiddens):
+        """
+        Handle forward for each training case (distributed, single gpu, etc...)
+        :param batch:
+        :param batch_idx:
+        :return:
+        """
+        # ---------------
+        # FORWARD
+        # ---------------
+        args = self.build_train_args(batch, batch_idx, opt_idx, hiddens)
+
         # distributed forward
-        if self.use_ddp or self.use_ddp2 or self.use_dp:
-            output = self.accelerator_backend.training_step(args)
-
-        # horovod
-        elif self.use_horovod:
-            output = self.accelerator_backend.training_step(args)
-
-        # single GPU forward
-        elif self.use_single_gpu:
-            output = self.accelerator_backend.training_step(args)
-
-        # TPU support
-        elif self.use_tpu:
-            output = self.accelerator_backend.training_step(args)
-
-        # CPU forward
-        else:
         output = self.accelerator_backend.training_step(args)
 
         is_result_obj = isinstance(output, Result)