Initializing a vectorized prior with multiple different distributions?

Questions
2

Hi,
You should be able to just do alpha = Normal('alpha', mu=np.array([-1., 1.]), sigma=1, shape=2)