Are you a book person? A give-me-a-sample-that-I-can-take-apart person? A video tutorial watcher? Someone who loves to parse deep into the technical docs? Or some combination of all of the above?
What has worked for you in learning TensorFlow, Machine Learning, or indeed anything?
For a scientific topic like Machine Learning (or any academic topic from fluid dynamics to literary linguistics), my approach is usually Full-length video courses + TextBooks.
For a practical platform like TensorFlow I tend to start (phase 1) with a couple (2-5) of video tutorials for a high level overview. I then move to (phase 2) written tutorials that show code samples and explain the code. The more samples and the smoother the progression from simple to more complex the better. After a couple (again 2-5) of tutorials I move to (phase 3) trying to implement a few things relying on documentation (official guide/API docs).
Books sometimes can be part of phase 2 as some books are basically written as very good tutorials (like these here).
For me, a book is the most consistent way as you get a more academic knowledge. However, it’s the slowest and should be coupled by checking examples, best practises and especially by rolling up your sleeves and starting coding and finding your way with examples and Kaggle competitions
Haha thanks! After Aurelien’s book, mine will feel basic! 
Excellent choice 
I also highly recommend for beginner, intermediate and advanced users:
) by @fchollet
Another excellent book:
From the publisher:
Deep Learning Design Patterns distills models from the latest research papers into practical design patterns applicable to enterprise AI projects. Using diagrams, code samples, and easy-to-understand language, Google Cloud AI expert Andrew Ferlitsch shares insights from state-of-the-art neural networks.
Another great resource:
@Sayak_Paul has curated list for resources to learn TF2.0 . Here it’s: GitHub - sayakpaul/TF-2.0-Hacks: Contains my explorations of TensorFlow 2.x