See PR for reference
Originally posted by @ricardoV94 in #5437 (comment)
mu=[0, 0]
Technically, it's incorrect to specify a scalar mu for the multivariate normal, although for backwards compatibility we reshape mu behind the scenes
Reply via ReviewNB