Randomizing array indices

I mentioned in my last post that I’m working my way through Practical Deep Learning by Ronald Kneusel. The math so far has been straightforward, and the code is written in Python, which I find easy to follow. “Easy to follow” does not mean “correct” or “to my liking.” Some of the code in the book simply doesn’t work, and I’ve had to make changes here and there to get it to run. It’s possible that some of the APIs the book relies on have changed their calling conventions, but I also found a pretty obvious off-by-one error. And I’m only three chapters in. This post, though, is about the “not to my liking” aspect of the code. I feel a little uncomfortable doing this, but I’ve been reading and writing data analysis code in Python for 15–20 years, so I’m going to put my discomfort to the side. One of the examples used in the book is the well-known iris measurement dataset. There are 150 records in the dataset; here are the first ten: Index Sepallength Sepalwidth Petallength Petalwidth…