Hi everyone,
I’m going through the forward propagation section in the coffee roasting model example, and I noticed what seems like a typo or error in one of the slides.
The activation vector a1 for the first hidden layer is defined as having 3 elements: a1 = np.array([a1_1, a1_2, a1_3]).
However, the weight vector w2_1 for the output layer is shown as np.array([-7, 8]), which only has 2 elements. The slide then attempts a dot product: z2_1 = np.dot(w2_1, a1) + b2_1.
This would cause a dimension mismatch error since you can’t dot product a 2-element vector with a 3-element vector. It looks like w2_1 might be missing a third value (perhaps it’s intended to be something like [-7, 8, some_value]).
Here’s the slide for reference.
Is this a known issue, or am I misunderstanding something? Thanks for any clarification!
Best,
Tejas Bhosale
