Vectors of vector parameters in Turing.jl?

General Usage
3

The error I was getting was because I was calling

sample(model,NUTS(0.8,500))

instead of

sample(model,NUTS(500,0.8))

So, I can basically say that yes it work to do the following:

    nationcoefs ~ MvNormal(zeros(Float64,nc),2.0)
    regioncoefs = fill(nationcoefs,nregions)
    for i in 1:length(regioncoefs)
        regioncoefs[i] ~ MvNormal(nationcoefs,2.0)
    end

It’s important to avoid doing something like fill(0.0,nregions) because it screws up the type system when Turing starts to use ReverseDiff types or ForwardDiff types etc.