C2M1_Assignment unittests.exercise_1(FlexibleCNN) error

Course Q&A Professional Certificate:PyTorch for Deep Learning PyTorch: Techniques and Ecosystem Tools
,
1 
Failed test case: Error during classifier checks: tuple index out of range.
            Check that you are creating the classifier correctly in the _create_classifier() method.
            Check that self.classifier uses a Sequential container with: dropout, linear, relu, linear layers..
Expected: Classifier to be constructed correctly.
Got: Error

Yet classifier is exactly as expected:
FlexibleCNN classifier architecture:
Sequential(
  (0): Dropout(p=0.5, inplace=False)
  (1): Linear(in_features=1024, out_features=128, bias=True)
  (2): ReLU()
  (3): Dropout(p=0.5, inplace=False)
  (4): Linear(in_features=128, out_features=2, bias=True)
)

Please help!
2

Hi @4eburek,

I’m going to send you a direct message regarding this.

3

Hi! I also got the same issue. What should I do? Thank you!

4

Hi @ruhiyahfw,

Sending you a direct message as well.

5

hi,

I got the same error and I am getting 38/40 as score because of this.

6

I have the same problem. I spent half a day on it and gave up. Couldn’t solve it. I rebooted, killed the kernel, deleted the file, uploaded it again. Not to mention how many times I checked the code :)))) What did they reply to you in the direct? :wink:

1 Like
7

Hi,

I also encountered the same problem, I check that my implementation and the expected output already matched but it failed the unittest.

Thank you

1 Like
8

I’ve got the same error as mentioned above. The output looks correct, but I fail the unit test.

"Failed test case: Error during classifier checks: tuple index out of range. Check that you are creating the classifier correctly in the _create_classifier() method. Check that self.classifier uses a Sequential container with: dropout, linear, relu, linear layers.. Expected: Classifier to be constructed correctly. Got: Error”

1 Like
9

I have got same the error.in create_classifier suggest 5 layers but in unittest_utils.py require 4 layers:
expected_classifier_params = [
(“Dropout”, {“p”: 0.1, “inplace”: False}),
(“Linear”, {“in_features”: 2048, “out_features”: 64, “bias”: True}),
(“ReLU”, {“inplace”: False}),
(“Linear”, {“in_features”: 64, “out_features”: 2, “bias”: True}),
]
what should I do ? remove dropout in assignment or add a dropout in unittest?

1 Like
10

Hi all,

We are working on the fix. It is very likely that your implementation is incorrect. And from what has been observed, it is likely due to unclear exercise and unit test instructions.

For example, the unit tests mentions:

While the test, as of now, ambiguously mentions the layer names you need, it does not specific how much of each or in what order you need those. We are working on fixing this, to make the test more clear.

Please follow the exercise and especially, the code cell instructions to be able to correctly implement the _create_classifier part, as they specify what you need to do at each layer.

Best,
Mubsi

11

We have fixed the test for exercise 1 (for the _create_classifier() method).
Now you should be able to get the full score if your implementation is correct.

Though, in some of the implementations there are some mistakes such as: a dopout layer is missing, or some argument should be called with self. but it is not, etc.

Please be sure to follow the exercise instructions, code comments and look at additional hints if needed to understand how and what needs to be implemented.

Best,
Julian