is there an efficient way to compute row sums? at the moment I am doing
(map core/sum (core/rows x))
yes! mv the matrix with a vector of ones.
excellent! must try harder to think in linalg instead of loops