Awesome content but one big flaw:
After 2 months using numpy to build neural networks (since course 1 of the specialization), this course briefly touches TensorFlow for around 2 hours.
I feel like we should at least do everything we did with numpy (l2 regularization, drop out, 2 layer nn, deep nn, etc) once again using TensorFlow.
It’s a good point that Course 2 just gives us a fairly brief introduction to TF. But it will be used more in Course 4 and Course 5, so you’ll continue to learn more about TF and Keras as you proceed through the courses. You can also check the TF online documentation and tutorials if you are curious about any of the topics you mention like regularization.