Well, after trying to figure it out for a few hours, I am here for help. The error I get is the following:
I did the dot product between the slice and W, I added all the values in “s” together with np.sum, and finally got Z by adding float(b).
I tried printing my results, and it seems that when this is casted: conv_single_step_test(conv_single_step)
the parameters b and W changed … could it be? Or where could my mistake be ?