Course-2, Week-1, Initialization Lab assignment issue

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

Hello!
The random initialization values returned after ‘random initialization’ in Exercise-2 does not match the results in the corresponding test. As a result, the test fails.

Did anyone have this issue?

Thanks for looking at my issue even if you are not able to help !!

-Param

2

Hi @Parama_Dutta,

Looking at the code segment relating to your question, I think you need to check the following:

  • Shape parameters given to the np.random.randn() call for the weight matrix
  • Whether the result of the np.random.randn() call is multiplied by 10
  • Shape parameters given to the np.zeros() call for the bias vector

Please see if these are correctly set as shown in the comment for the function.

3

The most common issue is using the wrong PRNG function. Note that rand is not the same thing as randn. That little n on the end there makes a big difference. :nerd_face:

4

Thanks. I was not paying attention to the rand->n function!

5

You guys rock man! That’s right!!