Tensorflow : Practice to Production using Kubernetes
Submitted by Vibhav Bobade (@vibhavbobade) on Tuesday, 5 June 2018
Technical level: Intermediate Status: Submitted
This session will move on from the practice datasets and models kept in a folder to moving Tensorflow into Production in multiple use cases.
When it comes to testing TF models it is relatively easy, but it is always a pain to move them into production.
We are going to discuss Data-Driven apps, Tensorflow Serving and Distributed Tensorflow Architectures (using Kubernetes)
This should help understand the challenges which we might face during production as well as some other challenges while changing the models in production. (with examples)
At the end of the talk users would have a decent idea about Systems Design with Machine Learning.
Vibhav is an enthusiastic individual who works as a Software Maintenence Engineer at Red Hat and likes to work on Data Driven Apps on the side. Areas of Interest : Distributed Systems, Data Driven Apps