LoginSign Up
← Back to all sparks
S

stacks

ANALYTICS
Velocity0.0

Model stacking and ensembling for the tidymodels framework

github.com/tidymodels/stacks ↗

Model stacking in tidymodels, quietly migrating off foreach and onto future

tidymodelsensemblingparallel-processingfuture-frameworkr-package
◆Current state
stacks builds ensembles from tidymodels tuning results, and its release history is dominated by one long project: replacing foreach-based parallelism with the future framework. That transition completed in 1.1.0, where foreach backends began being ignored with a warning and the minimum R version rose to 4.1. Releases are infrequent and small, with the most recent being a CRAN re-submission rather than a change.
◆Where it's heading
The package is mature and its remaining work is compatibility rather than capability — tracking the parallelism story across tidymodels, keeping object sizes sane after butchering and reloading, and staying aligned with recipes deprecations. The augment() method added for vetiver compatibility shows the same instinct: fit into the surrounding ecosystem rather than grow independently of it. Nothing in the visible history suggests new ensembling methods are being pursued.
◆Prediction
With the future migration finished, the next release is most likely maintenance keeping pace with tune and recipes rather than anything users would notice.

◆Recent moves

  1. 1y ago

    stacks 1.1.1 re-released to clear a CRAN check note

    A re-release of 1.1.0 to clear a CRAN note after a check failure was resolved. No change to the package itself.

    View source ↗
  2. 1y ago

    stacks 1.1.0 completes the move to future-based parallelism

    Parallel processing moves fully to the future framework, with foreach-registered backends now ignored and warned about, and the minimum R version rises to 4.1. It completes the migration started in 1.0.4 and is the one recent release with a real upgrade consequence.

    View source ↗
  3. 2y ago

    stacks 1.0.5 fixes butchered stack size inflation

    Fixes object-size inflation when a butchered model stack is saved and reloaded, plus a type-checking bug in add_candidates(name). Narrow bug fixes.

    View source ↗
  4. 2y ago

    stacks 1.0.4 introduces future-based parallel processing

    Introduces future-based parallel processing alongside the existing foreach support, with an explicit warning that foreach will enter a deprecation cycle. The start of the transition that 1.1.0 finished.

    View source ↗
  5. 2y ago

    stacks 1.0.3 clears recipes deprecations and a type-check bug

    Alt-text refinement, updated example objects to clear recipes deprecation warnings, a blend_predictions(mixture) type-check fix, and a CRAN documentation-alias notice. Pure maintenance.

    View source ↗
  6. 3y ago

    stacks 1.0.2 adds an augment() method for vetiver compatibility

    Adds an augment() method for model_stack objects for vetiver compatibility, converts the tree_frogs example variables to factors, fixes mgcv formula errors, and cuts evaluation time and memory during stacking. The vetiver method is the notable part, tying stacks into the deployment side of tidymodels.

    View source ↗