> the concurrency model of these ML libraries is not dependent on a CPU

Gotcha, my bad.

> Please at least learn some fundamentals about NNs and ML frameworks before contributing a long, uninformed response.

My apologies. It's not like I enjoy wasting people's time, starting with mine. It was an honest mistake, point taken, and thanks for explaining.

If you have a somewhat 'definitive' book or chapter/resource to share on "understanding concurrency"... please do so. I'm eager to learn, and I'd wager many reading such a thread.