Get list of parameters from Turing model

Specific Domains Probabilistic Programming
7

Nice, this looks so much less hacky than going through sampling (as I proposed earlier)!

Now, curiously, the two approaches give different parameter orders.

julia> DynamicPPL.syms(DynamicPPL.VarInfo(m))
(:σ, :μ)

julia> ch = sample(m,Prior(),1); ch.name_map.parameters
2-element Vector{Symbol}:
 :μ
 :σ

Which of the two orders is relevant for the parameter initialization in your initial post?