Resnet50 Comparator Error

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

Hi,
in the Resnet Assignment Exercise 3, after passing Exercise 1 and 2, I got an error in the following line
comparator(summary(model), ResNet50_summary)

with an error message,
Test failed
Expected value
[‘Conv2D’, (None, 2, 2, 512), 524800, ‘valid’, ‘linear’, ‘GlorotUniform’]
does not match the input value:
[‘Conv2D’, (None, 4, 4, 256), 262400, ‘valid’, ‘linear’, ‘RandomUniform’]

After looking at some posts on this site, I suspect that this may because of the ordering of the NN when adding the skip-connection layer to the model (Add()[…]). I tried to swap the ordering in the identity_block or convolutional_block, but so far it still does not work and could not pass the test. Do you have any suggestions for this?

Regards,
Leonardo

2

Hi leonardoedgar,

Did you manage to resolve this? If not, could you send me your code by means of a direct message, so I can have a look?

3

Have you found the solution to this problem? I’m facing it too. Could you help me?

Thanks

4

Hi salwyn13,

Yes, it was resolved by checking the number of identity_blocks in stage 4. Hope this works for you too.

1 Like
5

is this a common error?

6

This thread is four months old. Please start a new thread so the mentors can find it more easily.