Feature
In situations when running very large models with tune_race_anova in parallel using plan(multisession), I often encounter the following error:
Warning: Caught FutureInterruptError. Canceling all iterations ...
Error: Future (‘tune-grid-3’) of class MultisessionFuture interrupted, while running on ‘localhost’ (pid 1837)
This often happens after 12+ hours of execution, which is a really huge waste of time.
Is there any way that tune_race_anova could save interim results to hard disk, so when these errors happen, I can at least have the interim grid results to pass to the function when I try again?
Feature
In situations when running very large models with
tune_race_anovain parallel usingplan(multisession), I often encounter the following error:This often happens after 12+ hours of execution, which is a really huge waste of time.
Is there any way that
tune_race_anovacould save interim results to hard disk, so when these errors happen, I can at least have the interim grid results to pass to the function when I try again?