Changelog

feat: The design of batchmark() can now include parameter settings.
feat: reduceResultsBatchmark gains argument fun which is passed on to batchtools::reduceResultsList. Useful for deleting model data to avoid running out of memory. Thanks to Toby Dylan Hocking @tdhock for the PR (https://github.com/mlr-org/mlr3batchmark/issues/18).
docs: A warning is now given when the loaded mlr3 version differs from the mlr3 version stored in the trained learners
feat: support marshaling

feat: mlr3batchmark now depends on package batchtools to avoid having to load batchtools explicitly.

mlr3batchmark (development version)