The loss is therotically negative log-likelihood, and the likelihood is computed as {1/sqrt(2pi)}^n exp(-z^Tz / 2) Jac.
=> The negative log-likelihood: n/2 * log(2pi) + z^Tz / 2 - logJac
The likelihood can be itself over 1 anyway.
Because n/2 * log(2pi) aren't computed for the actual computing. (It is constant). The loss is lower than the real negative log-likelihood. Refer to loss in fastflow.py.