About week1 notes pdf p.50

On P.50, the lecturer demonstrate a forward propagation using the code below:

x = np.array([[200.0, 17.0]])
layer_1 = Dense(units=3, activation="sigmoid")
a1 = layer_1(x)
layer_2 = Dense(units=1, activation="sigmoid")
a2 = layer_2(a1)

I am just confused without training the model, how does tensorflow knows which parameter to use?

The code on next page (p.51) seems to make more sense to me. It trains the model using the line model.fit(x,y) before predicting new values.

Hi @empheart

In the lecture videos further on, you will find Prof. Ng explains in detail what is going on under the hood. So just hold your thought for now.

1 Like