C1M2 Graded Lab Grading Confusion

Course Q&A Agentic AI
,
1

  • Issues: The unit tests for generate_draft(), reflect_on_draft(), and revise_draft() all show All tests passed! in green fonts, but the autograder shows 0/30 earned. Since this assignment is mostly about writing prompts, it is confusing that I could not pass even one test case in the auto-grader, while the unit tests all pass.

  • Platform: DeepLearning.AI

@Deepti_Prasad

C1M3 Grading Confusion with Uninformative Error Messages
C1M2_Assignment - Errors on Submit
2

I mentioned in my response to take a screenshot of your submission grader feedback which will provide information about why you failed your assignment.

Also remember passing all unittest doesn’t confirm your codes are as per the grader.

An autograder check if you have not deleted, edited anything outside code markers ###START AND END CODE HERE###

Then checks if you have not missed or mixed any implementation methods autograder is looking for.

autograder also fails if your assignment name is changed. There are many reasons, to know what is the reason, first step is to check information provided on your grader submission. Next if that’s is not clear, a mentor reviews your codes (learner is advised to send code screenshots by personal DM). Finally after mentor reviews, they will reply on your created topic, with explanation on how you can debug or correct your codes if any correction(s) required.

3

Thank you for your prompt reply. I can confirm that the filename is not changed, and here is the output of the auto-grader:

image
image775Ă—650 60.7 KB

4

did you read the instructions I gave in the thread you first commented?

is your code aligned in the way it is mentioned here in my comments

If it’s aligned like how I explained in my post comment, and yet you have not passed grader. then I probably need to review your codes, you can send me codes screenshot by clicking on my name and then message.

5

@SupportVectorMachine

your prompts in all the exercises are very vague and concised. (Read my comment response! to the linked comment I shared here again. read how the prompt is written, and notice the position of string object topic how it is placed.

Same need to be followed in exercise 2 and 3, improve your prompt instructions, place f-string in right positions. You also have placed output string object before the prompt in the second part of exercise 3 for reflection.

Hint: Refer the requirements points, when instruction is mentioned to write an essay topic, prompt should include describing of how and what essay context should be.

For example if I am teacher telling my student
write essay on topic??

the next question or doubt students would have what topic ? or how should I write??.

so I tell them

prompt= f """ Write a detail essay topic on how you plan to spend your Christmas holiday this year.  The draft need to have proper introduction, description and summary(something like this)

then here👇🏻 mention the(don't! include this line, it is only for your understanding)
topic:
{topic}
"""

all your exercises need to follow the same pattern of prompt description and most important include the input and output instructions mentioned (don’t go for shortcuts here) as it will again cause the grader failure.

6

It is funny, as I passed the same prompt and now I passed with 30/30. There seems to be some bugs in your LLMaaJ auto-grader running in the background.

image
image637Ă—537 23.3 KB

2 Likes
7

Hi @enino @a-zarta can we have a look here, please?

8

I have tried to replicate but was not able to do so. The fact that it failed and then it worked makes me think that the notebook was not saved before submitting. So unit tests passed but then the grader got the unedited version of the notebook, then the autosave kicked in and next submission worked.

1 Like