I got the expected outputs as mentioned in Jupyter notebook for both Case 1: Stride of 1 and Case 2 : Stride of 2. It shows “all tests passed” for Stride 1 but I can’t see the message ‘All tests passed/failed’ for Case 2 stride of 2 and I get the expected outputs as mentioned in the notebook.
Once I submit the notebook I get a score of 75 and it says:
“Tests failed on 1 cell(s)! These tests could be hidden. Please check your submission.”
Are you sure you checked all the values in the pool_forward stride = 2 test case? That one does not include a message about whether the test passed or not, but all the values need to agree, not just most of them. The first rows agree even if you did not implement the stride correctly.
The test cases for conv_forward in the notebook also miss the error if you omit the stride.
There is one other mistake that is not caught by the tests in the notebook, which is coercing the output type of conv_single_step to float32.
Thank you for the clarification. I have been checking the issue at the stride 2 test case as there was no confirmation that ‘all tests passed’. The actual issue was with conv_forward function. I did not scale the filter position with stride in the computation and it still showed ‘all tests passed’. After correcting it and resubmitting the grader shows 100.