Add an example for local mode deployment of models

If there is an existing model it can be deployed locally
using local mode. This feature is already present but there
is no example in the README.

Author: Ignacio Quintero

README.rst

@@ -192,6 +192,24 @@ instance type.
     mxnet_estimator.delete_endpoint()
 
 
+If you have an existing model and would like to deploy it locally you can do that as well. Make sure
+you don't pass a real sagemaker_session to the Model constructor.
+
+.. code:: python
+
+    from sagemaker.mxnet import MXNetModel
+
+    s3_model = MXNetModel(model_data='s3://mybucket/my_model.tar.gz', role='SageMakerRole',
+                          entry_point=mxnet_model.entry_point, image=mxnet_model.image)
+
+    predictor = s3_model.deploy(initial_instance_count=1, instance_type='local')
+    data = numpy.zeros(shape=(1, 1, 28, 28))
+    predictor.predict(data)
+
+    # Tear down the endpoint container
+    predictor.delete_endpoint()
+
+
 For detailed examples of running docker in local mode, see:
 
 - `TensorFlow local mode example notebook <https://github.com/awslabs/amazon-sagemaker-examples/blob/master/sagemaker-python-sdk/tensorflow_distributed_mnist/tensorflow_local_mode_mnist.ipynb>`__.