Coursera NLP with Tensorflow Assignment 4

[Mojad1]

I have passed all test cases and even my accuracy is above 80 . Even then the grader shows me 25%. However, in my training accuracy it shows 25/25 but in every other topic it shows 0 and also corpus not defined. I have ran it like 15 times . What;s wrong ?

My answers match all the expected answers and also the test cases as well

Btw, corpus is a variable already defined in the notebook’s frozen funtions , I did not even define it . Every thing runs perfectly in the Jupyter nb but soemhow fails in the grader.

Is the server down or something ?

Pls help

It shows this for every metric except training. For training it shows 25/25

[Mojad1]

My course is NLP with tensorflow week 4

[TMosh]

I will move the thread to the forum area for that course.

Note: The grader tests your code using entirely different tests. Passing the tests in the notebook does not prove your code is perfect.

For example:
Be very careful to check that the code you added does not make incorrect use of global variables. The grader may not provide the same variables as are used in the assignment notebook.

[Deepti_Prasad]

hi @Mojad1

Please click on my name and send me screenshot of the fit_vectorizer grade function codes. Don’t post codes here as it is against community guidelines.

I probably doubt your codes might be right but when you are submitting the changes you must have not got updated. if this is the issue, then clear your kernel output, reconnect the kernel, run the cell from beginning till end one by one then submit (if you are confident of your codes being correct)

Regards

DP

[Mojad1]

Alright, thank you.

[Mojad1]

Hello, thank you for reaching out. I will send the you the code via DM.However, the screenshot I displayed is not a code rather grader’s output. Also, I have the same problem with all other functions except one that did not use corpus variable i.e Training function which is fine 25/25.

Regards,

Mojad

[Deepti_Prasad]

Hi @Mojad1

While defining object, instructions given are

So you only need to make sure the parameters for lower punctarion and tagged =True is added and remove the other parameters. make sure once you do the correction, always clear the kernel output and re-run the codes incase you are facing grader failure.

Regards

DP

[Mojad1]

Alright, so I should standardize and ragged(True) as well right ? I’ll make changes and update you.

[Mojad1]

1. Restarted the kernel and cleared output of all cells

2. Made the relevant changes

But still the same output by grader. 0 for all graded funcs except for training func(25/25)

[Deepti_Prasad]

as shared by you in DM that your other grade function codes also have scored 0, I need to go through your other grade functions codes too.

I will check it in sometime by then kindly move to the next week. I will reply back once I have checked your complete codes.

[Mojad1]

Sure, thanks for your time .

[Deepti_Prasad]

ok @Mojad1

I noticed in one of the grade function you have removed/deleted the given argument part, so probably that’s why autograder is failing your assignment submission.

I highly recommend you to get a fresh copy, and restart the assignment. While you do make sure to only replace the None place between the markers ###START AND END CODE HERE###. Anything outside these markers shouldn’t be edited/deleted/added with any new codes.

Other than that for pad sequence, remove data type parameter

in your create model, for embedding layer, you are suppose to mention the argument variable for output dim. Same goes for the lstm layer, do not put direct numerical values.

I hope you know how to get a fresh copy of assignment. Let me know if issue still persist.

Regards

DP

[Mojad1]

Yes, I do know how to request a new copy.

I will make the necesarry changes and get back to you.

Thanks for your help.

[Mojad1]

After making the changes you suggested. It gave me two grading errors and the third time finally worked.

Thank you so much for your guidance.

Regards,

Mojad