C4_W2_Check the padding and strides

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

I used strides = (1,1) and padding= “valid” for first and third and “same” for second and the out put gave me AssertionError Check the padding and strides
I read old topic in this week n found that the second layer filter =f not F2 but it also didn’t solve the problem

2

You might read that about kernel_size. The filter for the second component is F2.

3

Note that there are two assignments in C4 W2: Residual Networks and Transfer Learning with MobilNet. It sounds like you are probably asking about the Residual Networks assignment but it would be nice to be clear about that.

The next step would be to show us the actual error output that you are getting.

4

I’m asking about Residual Networks and here is the error

image
image1198×455 32.1 KB

5

I’m waiting you dear

6

Sorry I am UTC -8 so I was asleep until quite recently. So the output values are incorrect. This whole assignment is an excruciating exercise in proofreading. You have to check carefully all the parameters. The other common error here is how the shortcut layer is handled: are you sure you used the saved X_shortcut value as the input in the appropriate place. Although that mistake is more common on the convolutional_block.

7

Really I have checked the code line by line many time but I didn’t see any thing wrong can I sent you my code to check it ?

8

I believe I already send you a DM about how to do that. I will try again just in case.

9

I have sent

10

You have not followed the instructions correctly for the Second Component of the main path. The instructions tell you to use a filter size (kernel size) of (f, f), but you are using kernel_size = 1.

2 Likes