Wk 2: How to determine size of W while initializing

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

Hi,

When we are initialising W as a vector of zeroes while starting with model creation, how do we know how big W should be? We had set the size to 2 earlier in the assignment but what should be the approach when collating all intermediate functions to create overall algorithm?

2

I think I understood this. Each pixel is treated as a feature and hence 3*(number of pixels)^2 Ws are also reqd corresponding to them