This is the second lesson in the fast.ai deep learning course, which has been very popular on the internet. The course is completely free on their website. So, this course is taught primarily by Jeremy Howard, a popular figure in the deep learning and machine learning community.
I know for a fact that they taught this course last year - that's when it was the first run of it. And they probably had it going for this year as well because it caught well to the open public - which is specialized in machine learning. I dare to say this is not a beginner course.
And the approach, compared to that which is taught by Andrew Ng is a top down approach, not a bottom up approach. So, Jeremy starts with a high level perspective and as it the lessons proceed he gets into the necessary details (but not all of the granular details).
This is the second lesson and the student has already trained an algorithm to recognize things (in the first lesson) by running a couple of lines of Python code. Here Jeremy explains about the learning rate, how to use a certain script to find a good learning rate, and much more.
A few things I really liked about this course are that a lot of focus is put on the implementation and coding. Many topics that are new such as CycleGAN are not covered in many other courses available on coursera.
Downvoting a post can decrease pending rewards and make it less visible. Common reasons:
Submit
I havent gotten into those either, but I will someday.
Downvoting a post can decrease pending rewards and make it less visible. Common reasons:
Submit