The automatic grader of lab assignment of C1_W3 NLP course

I have a problem with NLP C1 Week3 grader. I fail the get_country and get_accuracy although I pass all tests within. I am stuck there. I have finished the 4 weeks and cannot move on because of this faulty grader? Can you please advise!
Assignment: Vector Space Models | Coursera

1 Like

You have probably made the most common mistake on this assignment, which is referencing the global variable word_embeddings from the body of your get_country function. What if the grader passes a different dictionary?

It is not a bug in the grader: your code is incorrect, but it happens to work in the notebook because that is the variable that is passed in those test cases. That mistake also causes the get_accuracy function to fail the grader, because it calls get_country.


I changed it from word_embedding to embedding only and it worked! Thanks a lot for your support and prompt reply!

Glad that my hint was enough for you to fix the problem! That’s an important point to remember: referencing global variables from the body of a function is always a mistake at least in these courses. There can be other ways to structure code where it is ok, but here they are very careful about modularity of functions.


Thank you so much

1 Like