Update lab 10 solutions

[Antoine-Bergerault]

In the lab 10 there is one function that is the perfect opportunity to use np.einsum, a method that turns out to be very useful in practice. It makes it easier to deal with linear operators with tensors (with shapes being like (N, C, H, W)) by removing the need of reshaping and being more efficient (less operations and with some frameworks easier parallelization on multiple devices).

So I thought it would be great to update the solution for build_distance_matrix.