Use stan's pathfinder variational inference algorithm — sampler_stan

This option uses stan's pathfinder algorithm via cmdstanr::cmdstanr for variational inference. Sampling is fast but yields potentially inaccurate results with unreliable uncertainty information. Currently, this sampler is recommended only for preview purposes, not for real-world inference.

Usage

sampler_stan_pathfinder(
  draws = 4000,
  num_paths = 4,
  single_path_draws = NULL,
  max_lbfgs_iters = NULL,
  num_elbo_draws = NULL,
  psis_resample = FALSE,
  seed = 0,
  refresh = 1000,
  show_messages = TRUE,
  show_exceptions = FALSE,
  ...
)

Arguments

draws: (positive integer) Number of draws to return after performing pareto smooted importance sampling (PSIS). This should be smaller than single_path_draws * num_paths (future versions of CmdStan will throw a warning).
num_paths: (positive integer) Number of single pathfinders to run.
single_path_draws: (positive integer) Number of draws a single pathfinder should return. The number of draws PSIS sampling samples from will be equal to single_path_draws * num_paths.
max_lbfgs_iters: (positive integer) The maximum number of iterations for LBFGS.
num_elbo_draws: (positive integer) Number of draws to make when calculating the ELBO of the approximation at each iteration of LBFGS.
psis_resample: (logical) Whether to perform pareto smoothed importance sampling. If TRUE, the number of draws returned will be equal to draws. If FALSE, the number of draws returned will be equal to single_path_draws * num_paths.
seed: (positive integer(s)) A seed for the (P)RNG to pass to CmdStan. In the case of multi-chain sampling the single seed will automatically be augmented by the the run (chain) ID so that each chain uses a different seed. The exception is the transformed data block, which defaults to using same seed for all chains so that the same data is generated for all chains if RNG functions are used. The only time seed should be specified as a vector (one element per chain) is if RNG functions are used in transformed data and the goal is to generate different data for each chain.
refresh: (non-negative integer) The number of iterations between printed screen updates. If refresh = 0, only error messages will be printed.
show_messages: (logical) When TRUE (the default), prints all output during the execution process, such as iteration numbers and elapsed times. If the output is silenced then the $output() method of the resulting fit object can be used to display the silenced messages.
show_exceptions: (logical) When TRUE (the default), prints all informational messages, for example rejection of the current proposal. Disable if you wish to silence these messages, but this is not usually recommended unless you are very confident that the model is correct up to numerical error. If the messages are silenced then the $output() method of the resulting fit object can be used to display the silenced messages.
...: Further arguments to pass to cmdstanr::cmdstanr.

Value

A list with settings for the pathfinder sampler.