W4-1_E7 : Variable db gives error

Course Q&A Deep Learning Specialization Neural Networks and Deep Learning
1

Hello,

Context: Week 4 (Part 1) Exercise 7.

Possible Issue: In spite of trying multiple options for parameter ‘db’, errors continue to exist.
Please guide what exactly is going wrong.

Options Tried:
# db = (1/m)*dZ # FAIL
# db = (1/m)*np.sum(dZ,axis=1, keepdims= True) #FAIL
# db = (1/m)*np.sum(dZ,axis=1, keepdims= False) #FAIL
# db = (1/m)*np.sum(dZ,axis=0, keepdims= True) #FAIL
db = (1/m)*np.sum(dZ,axis=0, keepdims= False)

Error:

db: [ 0.70219802 -0.79066637 0.66968699 -0.97357906]
Error: Wrong shape for variable 2.
Error: Wrong output for variable 2.
1 Tests passed
2 Tests failed

AssertionError Traceback (most recent call last)
in
6 print("db: " + str(t_db))
7
----> 8 linear_backward_test(linear_backward)

~/work/release/W4A1/public_tests.py in linear_backward_test(target)
385 ]
386
→ 387 multiple_test(test_cases, target)
388
389 def linear_activation_backward_test(target):

~/work/release/W4A1/test_utils.py in multiple_test(test_cases, target)
140 print(‘\033[92m’, success," Tests passed")
141 print(‘\033[91m’, len(test_cases) - success, " Tests failed")
→ 142 raise AssertionError(“Not all tests were passed for {}. Check your equations and avoid using global variables inside the function.”.format(target.name))
143

AssertionError: Not all tests were passed for linear_backward. Check your equations and avoid using global variables inside the function.

Screen Shot 2023-11-08 at 10.38.20 AM
Screen Shot 2023-11-08 at 10.38.20 AM1067×711 98.4 KB

Sincerely,
A

2

Instead of trying multiple things, read the instructions again and use your intuition to find out which one is correct. Of the equations you shared, one of them is correct.

3

Thanks Saif for response.
I got you. But, the correct solution should have no error.

In my view, calculations for variable ‘db’ should be column-wise, thus axis=0.

For db = (1/m)*np.sum(dZ,axis=0, keepdims= False):
Error Message: Screen Shot 2023-11-16 at 11.03.17 PM.png (attached)

For db = (1/m)*np.sum(dZ,axis=0, keepdims= True):

Screen Shot 2023-11-16 at 11.03.17 PM
Screen Shot 2023-11-16 at 11.03.17 PM901×734 98.5 KB

Screen Shot 2023-11-16 at 11.04.27 PM
Screen Shot 2023-11-16 at 11.04.27 PM984×728 103 KB

Error Message: Screen Shot 2023-11-16 at 11.04.27 PM.png (attached)

I also looked int the equation, and not sure how to pursue it forward…

Sincerely,
A

4

Please check this instruction from the notebook:

For the next exercise, you will need to remember that:

  • b is a matrix(np.ndarray) with 1 column and n rows, i.e: b = [[1.0], [2.0]] (remember that b is a constant)
  • np.sum performs a sum over the elements of a ndarray
  • axis=1 or axis=0 specify if the sum is carried out by rows or by columns respectively
  • keepdims specifies if the original dimensions of the matrix must be kept.
  • Look at the following example to clarify:
5

Hello Saif,

I am not aware of any notebook shared during the certification.
It would be great if you may share any course book.

I am currently referring to weekly .pdf files only.

“Look at the following example to clarify:”

  • I guess you were intending to post some examples as well.

Sincerely,
A