diff --git a/docs/source/introduction_guide.rst b/docs/source/introduction_guide.rst
index e6d8b33ce8..517fb10409 100644
--- a/docs/source/introduction_guide.rst
+++ b/docs/source/introduction_guide.rst
@@ -1,6 +1,7 @@
 .. testsetup:: *
 
     from pytorch_lightning.core.lightning import LightningModule
+    from pytorch_lightning.core.datamodule import LightningDataModule
     from pytorch_lightning.trainer.trainer import Trainer
 
 .. _introduction-guide:
@@ -259,9 +260,9 @@ In this case, it's better to group the full definition of a dataset into a `Data
 - Val dataloader(s)
 - Test dataloader(s)
 
-.. code-block:: python
+.. testcode:: python
 
-    class MyDataModule(pl.DataModule):
+    class MyDataModule(LightningDataModule):
 
         def __init__(self):
             super().__init__()
diff --git a/docs/source/lightning-module.rst b/docs/source/lightning-module.rst
index c6bf4453ba..6be28af555 100644
--- a/docs/source/lightning-module.rst
+++ b/docs/source/lightning-module.rst
@@ -51,7 +51,7 @@ Notice a few things.
 
         # or to init a new tensor
         new_x = torch.Tensor(2, 3)
-        new_x = new_x.type_as(x.type())
+        new_x = new_x.type_as(x)
 
 5.  There are no samplers for distributed, Lightning also does this for you.
 
diff --git a/docs/source/new-project.rst b/docs/source/new-project.rst
index 68b4f54afa..2ce86bd0cd 100644
--- a/docs/source/new-project.rst
+++ b/docs/source/new-project.rst
@@ -1,6 +1,7 @@
 .. testsetup:: *
 
     from pytorch_lightning.core.lightning import LightningModule
+    from pytorch_lightning.core.datamodule import LightningDataModule
     from pytorch_lightning.trainer.trainer import Trainer
     import os
     import torch
@@ -357,9 +358,9 @@ And the matching code:
 
 |
 
-.. code-block::
+.. testcode:: python
 
-    class MNISTDataModule(pl.LightningDataModule):
+    class MNISTDataModule(LightningDataModule):
 
         def __init__(self, batch_size=32):
             super().__init__()
@@ -407,7 +408,7 @@ over download/prepare/splitting data
 
 .. code-block:: python
 
-    class MyDataModule(pl.DataModule):
+    class MyDataModule(LightningDataModule):
 
         def prepare_data(self):
             # called only on 1 GPU
@@ -415,12 +416,12 @@ over download/prepare/splitting data
             tokenize()
             etc()
 
-         def setup(self):
+        def setup(self, stage=None):
             # called on every GPU (assigning state is OK)
             self.train = ...
             self.val = ...
 
-         def train_dataloader(self):
+        def train_dataloader(self):
             # do more...
             return self.train
 
@@ -432,7 +433,7 @@ First, define the information that you might need.
 
 .. code-block:: python
 
-    class MyDataModule(pl.DataModule):
+    class MyDataModule(LightningDataModule):
 
         def __init__(self):
             super().__init__()
@@ -444,7 +445,7 @@ First, define the information that you might need.
             tokenize()
             build_vocab()
 
-        def setup(self):
+        def setup(self, stage=None):
             vocab = load_vocab
             self.vocab_size = len(vocab)