C4W2A1 getting AssertionError: Wrong values when training=True

Course Q&A Deep Learning Specialization Convolutional Neural Networks
1

I keep getting AssertionError: Wrong values when training=True. I see in the changelog for the lab that training parameter was removed and thus I should not include them in the solution but this error pops up. Can anyone guide me in this? Thank you.

image
image1365×542 30.1 KB

2

The test works properly on the latest version of the notebook (just ran the test).

Please note that there is no training parameter in the signature for convolutional_block. So, don’t worry about setting the training state.

3

Hi could you help me understand what I am doing wrong? I do not want to post my code here.

4

Sure. Here’s the feedback after seeing your snippet:

  1. The bug is in this section of the function ##### SHORTCUT PATH ##### (≈2 lines)
  2. Pay attention to what you pass in as the input to the layer. See the provided reference image to understand from where input is taken.
3 Likes
5

It hasn’t been removed entirely - just from certain parts of the assignment where it wasn’t needed and caused much confusion.