Member-only story
TensorFlow Extended 101: (literally) Everything you need to know
Automate the workflow to build, validate, and deploy machine learning models.
Let’s be fair, most machine learning projects tend to start with manual workflows. This is fine, and nothing to worry about.
But, at some point, you’ll want to focus on new models instead of maintaining existing ones. That’s where TensorFlow Extended (TFX) shines.
TFX will help you lift your ML workflow to the next level: An automated end-to-end pipeline that allows machine learning to scale.
This article will guide you through the lifecycle of a machine learning pipeline with TFX.
What is TensorFlow Extended (TFX)?
If you are new to TensorFlow Extended, I recommend an article from one of my colleagues, Darren who wrote a great introduction to TFX.
Tensorflow Extended (TFX) is designed to build end-to-end machine learning pipelines.
The first time I read about TFX was in 2019 (the year it was officially released to the public). Admittedly, it was quite overwhelming.
Instead, by understanding the pieces step by step, you’ll get a clear and easy understanding of how TFX can be used.
I am sure you will arrive at a point where you feel TFX is incomplete. In that case, consider TFX’s version number still starts with zero (0.27.0). If you ever wanted to be part of something from the very beginning, TFX is your way to get involved.
TensorFlow Extended (TFX) isn’t something you get through in between two coffees. Reaping the rewards of TFX is like growing tomatoes, it is simple but it takes time 🍅.
I promise, by the end of this post you’ll ask yourself “Why didn’t I used TFX way earlier?”
