Well, this course is about the practical sides of coding, not theoretical. Lots of libraries, easy to switch to another. However, as I see, people are bored with the process of learning constantly. They do know that is is easy to learn new programming language, it is easier to learn new libraries. However, people can not adapt to the change because change is faster than BMW E30.
Well, if you are new to Python and Deep Learning, this course is suitable for you. Also, if you are programmer, it’ll make you switch faster.
MNIST is a perfect toy problem to understand Convolutional Neural Networks. No need to understand about the channels(RGB), doesn’t require complex Convolutional Neural Networks architecture. Lots of benefits.
CIFAR 10 is as well a toy dataset. Enthusiasts start to watch tutorials and think it is perfect easy. Story differs, students try to adjust themselves to complexity, enthusiasm escapes and nothing changes.
This course tries to minimize number of people who lose the enthusiasm.