Transitioning from pm.sample() to pm.fit()

[jessegrabowski]

In my (admittedly limited) experience, I’ve found RMSProp to be a good one to try first.