Week 2 assignment assertion error for variable w and b

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

I’m facing problem in assignment where i intialize variables w and b to zeros but after running the code it shows assertion error == float . What is the reason i’m using np.zeros(x)

2

Hey @Asad_Ullah_Wallana,
It looks like you have initialized b as 0. As mentioned in the function’s description, it’s supposed to be of type float. Now, look into what you can do to transform it into a float-type value, or better, initialize it as a float-type value. Let me know if this helps.

Cheers,
Elemento

3

The other thing to check is that you used np.zeros only for the initialization of w. Note that b is a scalar. If you initialize b with np.zeros, then the type is np.array, not float, so the assertion will still fail.

4

i have defined as np.zeros(x, dtype=float) and still same error of assertion

5

i think there is some problem with the sheet now the error has vanished and the new problem is data type not understood for w also it is showing me syntax error at dw and when i removed the dw and db from code it showed the same syntax error at cost

6

Also i want to ask another question that when i initialize b to zeros (x,1) then an error "data type not understood

7

The syntax error you can’t find is probably caused by a missing close paren on the previous line.

It looks like you didn’t understand what I said in my previous reply. Here is the same point expressed in a different way:

w is a vector, so you need to initialize it using np.zeros. The point is that the output of np.zeros is a numpy array containing floating point zeros of the requested dimensions.

b is a scalar value, so you cannot use np.zeros in that case. You just assign it a value with a simple python assignment statement. But there you need to keep in mind the point that Elemento made in his earlier reply: in python 0 and 0. are not the same thing: the first one is an integer (which will fail that assertion again) but the second one is a floating point value. See the difference?

8

You also need to be careful about the syntax of the np.zeros call. Here is the documentation, which I found by googling “numpy zeros”. Notice that it takes a “tuple” as the argument to give the shape that you want.

But as mentioned above, it is a mistake to use it in the case of initializing the bias value b.

9

thanks a lot, I understood your point and the problem is solved as well .