so, where is activation function or pooling layer?

[yottacto]

And in CNN how to set the conv_params like pad, stride, etc. Actually, i'm not sure how you converse convolution to matrix multiplication, it looks like there are no filters at all?