Apply suggestions from code review

Author: Borda

docs/source/introduction_guide.rst

@@ -285,8 +285,8 @@ However, if you have multiple optimizers use the matching parameters
 
     class LitMNIST(pl.LightningModule):
 
-      def configure_optimizers(self):
-        return Adam(self.generator(), lr=1e-3), Adam(self.discriminator(), lr=1e-3)
+        def configure_optimizers(self):
+            return Adam(self.generator(), lr=1e-3), Adam(self.discriminator(), lr=1e-3)
 
 Training step
 ^^^^^^^^^^^^^

pytorch_lightning/core/lightning.py

@@ -224,7 +224,7 @@ def training_step(self, batch, batch_idx, hiddens):
             The presented loss value in progress bar is smooth (average) over last values,
              so it differs from values set in train/validation step.
         """
-        warnings.warn('training_step must be implemented to be used with the Lightning Trainer')
+        warnings.warn('`training_step` must be implemented to be used with the Lightning Trainer')
 
     def training_end(self, *args, **kwargs):
         """
@@ -1076,8 +1076,7 @@ def configure_optimizers(self):
                   }
 
         """
-        warnings.warn('configure_optimizers must be implemented to be used with the '
-                      'Lightning Trainer')
+        warnings.warn('`configure_optimizers` must be implemented to be used with the Lightning Trainer')
 
     def optimizer_step(
             self,
@@ -1279,7 +1278,7 @@ def train_dataloader(self):
                     return loader
 
         """
-        warnings.warn('train_dataloader must be implemented to be used with the Lightning Trainer')
+        warnings.warn('`train_dataloader` must be implemented to be used with the Lightning Trainer')
 
     def tng_dataloader(self):  # todo: remove in v1.0.0
         """Implement a PyTorch DataLoader.