Tensorflow 2.0 was released in September 2019. Since the release, I have tried to follow a couple of resources trying to learn it. One problem I found during my learning is that these materials tend to focus on very high-level APIs without a detailed walkthrough on the lower-level building blocks. There is nothing wrong with quickly get hands dirty in few lines of code, but it is not enough when we try to tackle problems that are different in shapes and sizes compared to the tutorial examples. I wished there can be a slower but more of a from the ground up path to mastering the Tensorflow 2 eco-system.
The ideal target audiences for this writing are Tensorflow users with some understanding of neural networks and basic exposure to higher-level Tensorflow/Keras APIs but wanted to gain an intermediate mastering to enable more customization.
Due to my personal limit, the understanding and examples shown here can be error-prone and sometimes misleading(in a way that I did not realize), suggestions and corrections are greatly appreciated.