The graphics are a bit simple, but what a great introduction to this topic.  What I really like is the first part, where he introduces neural network concepts.  The Jupyter part is solid, but perhaps less interesting.  Of course if you are using PyCharm for ML, then I recommend that you watch Get started with TensorFlow's High-Level APIs. Its second part is about developing with TensorFlow and PyCharm.

TimeLine

0:00 About the old version of this talk

0:43 I start the video here.

1:10 How Neural Networks Concepts

5:45 Build a TensorFlow app in Jupyter

12:45 Training Parameters

14:40 Training the Model.

15:40 Overfitting

17:10 Saving and Loading Models

19:00 Conclusion and The Future

Editors Note: