here are the condition for Exercise 1-5 (tried to attach but max 1 media)
but in submission I got this error:
please help. I’ve restart kernel and clear output several times.
You can always restore the original version in the 3 dots on the top right corner of the Lab! But perhaps it might be a bug in your account. I see some cases in the Forum here. May be @chris.favila can also have a look at this!
Hello @Tiofilus_S,
The common thing in Exercise 3, 4 and 5 is that you provide a Tensorflow model for the autograder to check, and a valid Tensorflow model has a build attribute. The message says 'NoneType' object has no attribute 'build', so I believe somehow your submission was giving the autograder None objects instead of Tensorflow models.
The quick fix would be, as @gent.spah suggested, restore the original version, but after you have backup your current work. To backup, you may download the notebook to your computer, or you may just copy-and-paste your works to a text editor and save it.
With the original version, you may copy-and-paste your works back but make sure you only make changes in the areas bounded by these lines:
### START CODE HERE ###
### START CODE HERE ###
If you strictly follow these steps, the autograder should be able to see your Tensorflow models.
Cheers,
Raymond
1 Like
Hi @gent.spah & @rmwkwok , thank you for the information. I just tried to reset it to original (after download the answer) then tried to pasted back each answer, exactly only changing the codes in marked area. But the result still the same.
1 Like
I was able to get full marks by following the steps in my last response.
@chris.favila, is there anything you can do from your side? From @Tiofilus_S’s response, it seems that even making and only making changes as required can result in that the autograder consistently (in all 3 exercises) graded on unexpected (None) objects.
1 Like
Hi! Can you send me your latest notebook via direct message? You can click my name then the blue Message button. Thank you.
1 Like
alright, I’ll send
Hi! As per the submitted file, this is a problem with the grader being too strict. Your solution should be correct but for now, please avoid adding tf.keras.Input(shape=(...)) in your model definitions (ie. you can start with the Dense layer right away). If you remove those, you will get the perfect score. We’ll note this down for the next grader revision. Hope that’s okay. Thank you.
3 Likes
Done. Thank you very much, Chris.
Case closed.
1 Like
Great! Glad to hear that! Will close this topic now.