could someone please clarify in the building of neural network using Tensorflow fuction of forward propagation, how Z3 dimension is (6,2). should not it be (6,1)? I am waiting for the reply. Thank you
The way the dimensions are arranged there is the same as in Course 1. The first dimension is the number of output neurons (6 for the output layer here) and the second dimension is the number of samples (m). For the particular test case they give you, there are 2 samples. For any one sample, the output of the network will be 6 x 1. Also note that they have (purposely) not included the activation at the output layer there. It will be done internally by the loss function by using the from_logits argument (when we get to computing the cost).