It would be awesome if we can use `thermox.log_prob` and gradient ascent to fit a multivariate [Vasicek model](https://en.wikipedia.org/wiki/Vasicek_model) to some real data
It would be awesome if we can use
thermox.log_proband gradient ascent to fit a multivariate Vasicek model to some real data