Example of using TensorFlow with IBM Event Streams, to explain how to get started creating machine learning applications using the data you have on Kafka topics.
This is a simple demo app, not intended for production use (there is no error-handling at all, the ML model isn't hosted, etc.). The aim is to provide an easy-to-understand working example of how to integrate Kafka with TensorFlow.
I've explained how it all works in https://dalelane.co.uk/blog/?p=3924
It's written for Python 3. To run it:
pip install -r requirements.txt
config.envwith the properties of your cluster
run-step-0.shto set up your Kafka topics
run-step-1.shto train a machine learning model using data from a Kafka training topic
run-step-2.shto train a machine learning model and use it to classify an image
run-step-3.shto train a machine learning model and test it using data from a Kafka test topic
run-step-4.shto use a machine learning model to classify a stream of events on a Kafka topic
To wipe everything if you want to start again, run