Visual recognition with TensorFlow and OpenWhisk

Machine learning is currently all around us. From online shopping and reading recommendations to devices that recognize verbal commands and applications that reroute your driving instructions to account for traffic accidents, more and more people find machine learning an important part of their daily interactions. And going forward, everyday lives might soon be affected by cars that drive themselves. As a developer, it’s only natural to want to dive-in and understand how to use these technologies. But where should you start?

There is no shortage of learning libraries to choose from, but TensorFlow has emerged as one of the most popular. TensorFlow describes its library as “an open source machine learning framework for everyone.” Available for a wide range of languages, most recently TensorFlow is also now available in a JavaScript library and even for mobile platforms.

We’ve published a code pattern to help you get started with machine learning. It takes images from an Anzi Cozmo robot and then trains a model with those images to allow the robot to recognize further images of its environment. (But don’t worry if you don’t have this toy. You can use whatever pics you like.)

The code pattern also leverages the power of Kubernetes and serverless technologies on the IBM Cloud – a good introduction to those platforms as well! Try it out to determine how you can use these technologies together in your own environment.