Defines a list specifying the arguments passed to either
rstan::sampling()
or cmdstanr::sample()
. Custom settings can be supplied which override the
defaults.
Usage
stan_sampling_opts(
cores = getOption("mc.cores", 1L),
warmup = 250,
samples = 2000,
chains = 4,
control = list(),
save_warmup = FALSE,
seed = as.integer(runif(1, 1, 1e+08)),
future = FALSE,
max_execution_time = Inf,
backend = c("rstan", "cmdstanr"),
...
)
Arguments
- cores
Number of cores to use when executing the chains in parallel, which defaults to 1 but it is recommended to set the mc.cores option to be as many processors as the hardware and RAM allow (up to the number of chains).
- warmup
Numeric, defaults to 250. Number of warmup samples per chain.
- samples
Numeric, default 2000. Overall number of posterior samples. When using multiple chains iterations per chain is samples / chains.
- chains
Numeric, defaults to 4. Number of MCMC chains to use.
- control
List, defaults to empty. control parameters to pass to underlying
rstan
function. By defaultadapt_delta = 0.95
andmax_treedepth = 15
though these settings can be overwritten.- save_warmup
Logical, defaults to FALSE. Should warmup progress be saved.
- seed
Numeric, defaults uniform random number between 1 and 1e8. Seed of sampling process.
- future
Logical, defaults to
FALSE
. Should stan chains be run in parallel usingfuture
. This allows users to have chains fail gracefully (i.e when combined withmax_execution_time
). Should be combined with a call tofuture::plan()
.- max_execution_time
Numeric, defaults to Inf (seconds). If set wil kill off processing of each chain if not finished within the specified timeout. When more than 2 chains finish successfully estimates will still be returned. If less than 2 chains return within the allowed time then estimation will fail with an informative error.
- backend
Character string indicating the backend to use for fitting stan models. Supported arguments are "rstan" (default) or "cmdstanr".
- ...
Additional parameters to pass to
rstan::sampling()
.
Value
A list of arguments to pass to rstan::sampling()
or
[cmdstanr::sample().
Examples
stan_sampling_opts(samples = 2000)
#> $chains
#> [1] 4
#>
#> $save_warmup
#> [1] FALSE
#>
#> $seed
#> [1] 94353714
#>
#> $future
#> [1] FALSE
#>
#> $max_execution_time
#> [1] Inf
#>
#> $cores
#> [1] 1
#>
#> $warmup
#> [1] 250
#>
#> $control
#> $control$adapt_delta
#> [1] 0.95
#>
#> $control$max_treedepth
#> [1] 15
#>
#>
#> $iter
#> [1] 750
#>