Overview

This is a simple demo of using Cloud Foundry, Flask, Python 3, and TensorFlow.

Requirements

• You must have at minimum a Predix Free Tier account.
• You must have the Cloud Foundry CLI and can run cf login

Setup

git clone https://github.com/j12y/predix-tensorflow-example
cd predix-tensorflow-example

You'll want to change the application name to something unique just for you.

cf push --random-route

Results

When you push the app you should see something like the following.

App predix-tensorflow-example was started using this command `python app.py`

Showing health and status for app predix-tensorflow-example in org
***************@**.com / space tensorflow as ***************@**.com...
OK

requested state: started
instances: 1/1
usage: 1G x 1 instances
urls: predix-tensorflow-example-subordinal-lid.run.aws-usw02-pr.ice.predix.io
last uploaded: Fri Sep 1 15:14:37 UTC 2017
stack: cflinuxfs2
buildpack: https://github.com/cloudfoundry/buildpack-python.git

     state     since                    cpu      memory         disk
details
#0   running   2017-09-01 08:16:52 AM   162.6%   251.9M of 1G   487.5M of 1G

You can add https:// to the url listed and see your simple tensorflow app running.

Description

This is the basic example from Getting Started with TensorFlow.

https://www.tensorflow.org/get_started/get_started

When combined with the Predix SDK you can start pulling data from services like Predix Asset and Predix Time Series. For more information about the SDK see:

• https://github.com/PredixDev/predixpy

For more details on how Flask works, what runtime.txt, Procfiles, etc. are about see the basic flask app repository:

• https://github.com/j12y/predix-py3-flask

---