log-likelihood calculation: logdet(u) instead of logdet(v)

[gudovskiyd]

Hi @joeybose I wonder why do you use u (=data[0]) instead of v (=data[1]) in the logdet calculation?

log(z) = log p(v) - log det [(\partial / \partial v) proj_{\mu}(v)]

https://github.com/joeybose/HyperbolicNF/blob/e5c457ea472e18c0002b7ae7fc13fa9573e0bf3e/distributions/wrapped_normal.py#L51