A Practical Guide to Machine Learning with TensorFlow 2.0 & Keras# Deep Learning, Machine Learning, & Neural Networks Resources

Transcript from the "Deep Learning, Machine Learning, & Neural Networks Resources" Lesson

[00:00:00]

>> We're also gonna be relying on three books, Deep Learning with Python by François Chollet. He is actually the owner of Keras. Keras is one of the deep learning frameworks. It was a standalone framework designed precisely just to gonna create this high level API's for deep learning. And Keras originally was using Torch, TensorFlow and a couple of other deep learning frameworks on the backend, right?

[00:00:33] But right now Keras became part of the TensorFlow itself. So, right now you can just import Keras, for instance, by typing importtensorflow.keras. And interestingly enough, all the API architecture or all the API structure got preserved. So if you find any examples about Keras from which we're gonna use TensorFlow version 1, you can still use it because Keras syntax haven't changed at all.

[00:01:05] You will just need to instead of typing import keras, type input tensorflow.keras and that's it. All right, so Francois Chollet is the author of Keras and he's got this really nice book Deep Learning with Python, published about two years ago. But as I said, just because of the preservation of Keras APIs, all of the examples in this book can be used even with TensorFlow 2.0.

[00:01:28] Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow by Aurelien Geron is a really good book. And we will be using actually examples from the book in the last part of the course, when we'll be talking about reinforcement learning. Actors, robotics, all of that, so that's part of the reinforcement learning.

[00:01:49] And also if you want to know specifics of TensorFlow itself, Hands-On Neural Networks with TensorFlow 2.0 by Paolo Galeone is the really good book. It also explains the internals of TensorFlow 2.0, so are worth, and I put all the links to their GitHub repositories with examples. So for instance, if you click to author names, it will redirect you to GitHub where the corresponding source code which they used in their books will present so you can play.

[00:02:28] All of those are Python based, once again, like I haven't included a JavaScript source code here, but if you would like to play with Python code, that's the way to find it really interesting examples. If you just want to learn more about TensorFlow, tensorflow.org is the good way to start.

[00:02:47] So, it will tell you how to install TensorFlow. On tensorflow.org, it's under heavy load for some reason, but anyway, that's the way where you can start. TensorFlow allows you to have almost the whole ecosystem for machine learning. So with JavaScript, you can run codes, or create models, train the models and to do the inference right in your browser.

[00:03:16] But you can also create mobile and IoT mods and deploy them for instance, on Raspberry Pi's, or something like that, right? And, of course, for the production, if you working for the company, TensorFlow provides this whole model training, deployment ecosystem. And we probably gonna touch it a little bit at the end of this class.