Resume a lame() MCMC run from a checkpoint

For a fit started with lame(..., checkpoint_path = "X.rds") that terminated early (either via max_seconds or an external interruption), continues the chain from the most recent checkpoint. The implementation is pragmatic: it loads the saved RNG state and re-invokes lame() with the original call arguments. The result is a fresh fit that picks up where the previous one left off in the random-number stream; the underlying MCMC counter restarts at 1.

Usage

lame_resume(path, nscan_more = NULL, ..., .envir = NULL)

Arguments

path

Checkpoint file path (the one passed as checkpoint_path to lame()).

nscan_more

Optional integer, the new nscan for the continuation. If NULL, the original nscan is used.

...

Additional arguments forwarded to lame() (override the saved values).

.envir

Environment in which to evaluate the saved call's data arguments. Defaults to the caller's frame; used internally when lame() forwards a resume.

Value

A fitted lame object.

Details

Equivalent consolidated entry point. lame(resume_from = path, ...) short-circuits to this function with the user-supplied overrides forwarded; pass nscan = K on the resume call to request K additional stored draws. Both call shapes are supported. The lame_resume(path, ...) form is safer for nested calls: the consolidated form re-evaluates the saved lame() call in parent.frame(), which is the lame() frame whose required formal arguments (Y, Xdyad, etc.) were not supplied on the resume call. Prefer lame_resume(path, ...) when calling from inside other functions or from non-global scopes.

Use nscan_more = K to override the nscan value to K for the continuation. Other arguments can be overridden by passing them to ....

Examples

# \donttest{
ck <- tempfile(fileext = ".rds")
data(YX_bin_list)
# short run with very-aggressive max_seconds to force early termination
fit1 <- lame(YX_bin_list$Y, YX_bin_list$X, family = "binary", R = 0,
             nscan = 5000, burn = 50, odens = 5,
             checkpoint_path = ck, checkpoint_every = 50L,
             max_seconds = 0.5, verbose = FALSE)
#> Warning: `family` = "binary" but `Y` contains values other than 0/1.
#> ℹ `Y` will be thresholded to `1 * (Y > 0)`; if you meant counts, use "poisson",
#>   or "ordinal"/"normal" as appropriate.
if (isTRUE(fit1$terminated_early)) {
  fit2 <- lame_resume(ck, nscan_more = 200)
  dim(fit2$BETA)
}
#> Warning: `family` = "binary" but `Y` contains values other than 0/1.
#> ℹ `Y` will be thresholded to `1 * (Y > 0)`; if you meant counts, use "poisson",
#>   or "ordinal"/"normal" as appropriate.
#> [1] 40  4
# }