par_stan: Run stan chains in parallel

In kleinschmidt/phonetic-sup-unsup: Supervised/unsupervised adaptation analysis

Description

Uses 'mclapply' as backend, and 'sflist2stanfit' to convert to a single, multi-chain ‘stanfit' object. If you don’t supply the 'fit' parameter (previously compiled Stan model object) then it will compile one and re-use it for all the chains. Returns a single

Usage

par_stan(data, file, fit = stan(file, data = data, chains = 0), chains = 4,
  cores = chains, rng_seed = 1, ...)

Arguments

data	same as 'stan'
file	filename for a stan model file. ignored if 'fit' also specified.
fit	fitted 'stan' object. Overrides any provided 'file' argument.
chains	number of chains to run (default: 4).
cores	number of cores to use, which controls how many chains can be run simultaneously. (passed to 'mclapply', so see there for more information) (default: 'chains')
rng_seed	optional, passed to each chain. (default: 1).
...	additional arguments passed to 'stan' for each chain.

Value

A single 'stanfit' object with 'chains' chains.

kleinschmidt/phonetic-sup-unsup documentation built on May 20, 2019, 12:33 p.m.