Change headers in examples README

Change headers in examples/README.md, make the title "How to Run a Deep
Learning Job"

examples/README.md

@@ -18,7 +18,7 @@
 -->
 
 
-# AII Examples
+# How to Run a Deep Learning Job
 
 ## Introduction
 
@@ -30,9 +30,7 @@ The system supports training or evaluation with CNTK, TensorFlow, and other cust
 This guide assumes users have already installed and configured the system properly.
 
 
-## How to Run a Deep Learning Job
-
-### Custom Docker Image
+## Custom Docker Image
 
 The deep learning jobs will run in docker containers in the system. Docker images need to be prepared in advance. We provide base docker images with HDFS, cuda and cudnn support so that users can build their own custom docker images based on it.
 
@@ -57,7 +55,7 @@ docker push localhost:5000/aii.run.tensorflow
 The built image can be used in the system now.
 
 
-### Config File
+## Config File
 
 Users need to prepare a json config file to describe the details of jobs, here is its format:
 
@@ -103,7 +101,7 @@ Here's all the parameters for job config file:
 | `retryCount`                   | Integer, optional          | Job retry count, no less than 0          |
 
 
-### Runtime Environment
+## Runtime Environment
 
 All user jobs will run separately in docker containers using the docker image specified in config file. For a certain job, each task will run in one docker container. The allocation of docker containers are influenced by resources on each node, so all containers in one job may on one node or different nodes. It's easy for one task in a job running without communication. But for distributed deep learning jobs, some tasks must communicate with each other so they have to know other tasks' information. We export some environment variables in docker container so that users can access to runtime environment in their code.
 
@@ -130,7 +128,7 @@ Here's all the `AII` prefixed environment variables in runtime docker containers
 | AII_TASK_ROLE\_`$i`\_HOST_LIST     | Host list for `AII_TASK_ROLE_NO == $i`, comma separated `ip:port` string |
 
 
-### Deep Learning Job Example
+## Deep Learning Job Example
 
 Users can use the json config file to run deep learning jobs in docker environment, we use a distributed tensorflow job as an example:
 
@@ -175,7 +173,7 @@ Users can use the json config file to run deep learning jobs in docker environme
 ```
 
 
-### Job Submission
+## Job Submission
 
 1. Put the code and data on HDFS