Back when I took this Specialization, DeepLearning didn’t explicitly support learners attempting to run the exercises locally - too much possible variation. Those who wanted to do so were mostly on their own, though if you search these fora you can find some helpful, if somewhat stale, threads discussing it. In any case, it’s easy enough to create a dependency listing by adding code to one of the exercise notebooks…
$ pip freeze > requirements.txt
You should have an understanding of how to create and manage tailored environments using tools similar to conda before heading down this path.
ps: here’s a related thread that contains a link to another…