StanLogDensityProblems

StanLogDensityProblems implements the LogDensityProblems API for Stan models using BridgeStan. For easily benchmarking inference algorithms, StanLogDensityProblems also integrates with PosteriorDB.

For example, here we sample a Stan model from PosteriorDB using DynamicHMC:

julia> using DynamicHMC, PosteriorDB, Random, StanLogDensityProblems

julia> pdb = PosteriorDB.database()
PosteriorDatabase(...)

julia> post = PosteriorDB.posterior(pdb, "dogs-dogs")
Posterior: dogs-dogs

julia> prob = StanProblem(post, ".")
StanProblem: dogs_model

julia> rng = Random.default_rng();

julia> result = mcmc_with_warmup(rng, prob, 1_000; reporter=NoProgressReport());

julia> result.posterior_matrix
3×1000 Matrix{Float64}:
  1.27568    1.29648    1.32405   …   1.65451    1.3875     1.76917
 -0.327308  -0.305644  -0.303549     -0.273245  -0.387199  -0.333485
 -0.128237  -0.166094  -0.15898      -0.223369  -0.15062   -0.236186