Take a deep NN in figure-1 as an example,
![](https://i-blog.csdnimg.cn/blog_migrate/e257aed489aef6a8c8eb0af7a92bcaec.png)
- For a single training example x
Note the dimensions of these variables:
Concretely,
- For multiple (m) training examples.
Note the dimentions of these variables:
will be duplicated to
in NumPy automatically through Python broadcasting.
Concretely,
Note that we're using capital X, Z, A now which are all stacking values of m training examples together in columes. For example, here X looks like:
![](https://i-blog.csdnimg.cn/blog_migrate/997d4304ccb0b1401b6c1dc23298a397.png)
<end>