Hi, I got some strange feedback and a fail, but I think the code works.
Failed test case: model was not originally set to train for 10 epochs.
Expected:
10,
but got:
1.
As you can see the callback works and stops the training after 5 iterations. I set 10 epochs as the instructions stated it must end in under 10 iterations.
Please click my name and message your notebook as an attachment.
the loop was not breaking so see that inside the loop the code is correct or not
the last line is the output from model.evaluate - i wonder if that was the issue. If you look on the epoch count on the left you can see it stops after 5.
its stops due to accuracy has reached max limit ,
@joshctaylor
Your guess is correct. Calling model.evaluate produced the garbage feedback. Please remove the call to get the full score. The staff have been notified to revise the grading infrastructure. Thanks.
thanks - i figured that was it - it would be useful to guide us in the question on what needs to be in the output for the grader to work.
There are a lot of hints in the notebook.
That said, there are no notebook level unit tests. This mimics the tensorflow developer certificate exam by Google.
If you need anything specific, please provide the details on what exactly you are looking for. I’d be happy to forward your ask to the staff / ask them to get in touch with you.
I think just a paragraph or short video on how the grader is doing its job would help people to understand the task.
The staff have been informed about your ask.
Hi Josh! Welcome to the community and sorry for the inaccurate grader output! Ideally, the learners shouldn’t worry about the grader details. They can focus instead on solving the labs based on the info given in the lectures and the tips in the notebook. Nonetheless, there are a few bugs that slip and we’re very grateful that you reported this. This will be fixed so other learners won’t run into it. Thanks again!
