-
Notifications
You must be signed in to change notification settings - Fork 2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AutoML Pipeline – Integration to AutoML #15857
Comments
sebhrusen
added a commit
that referenced
this issue
Feb 15, 2024
* AutoML pipeline support * revert model type detection in AutoML SE support
mn-mikke
added a commit
that referenced
this issue
Feb 27, 2024
wendycwong
pushed a commit
that referenced
this issue
Feb 29, 2024
…16061) * cleanup legacy TE integration in ModelBuilder and AutoML * renamed ProcessingStepDefinition schema to PipelineStepDefinition for consistency * fixed NPE due to short-path in applyPipeline * addressed tomf suggestions * fixed serialization issue for frame trackers * fixed key leakage in multinode * fixed Py+R automl_preprocessing expectations * make all transformers Keyed for better serialization * fix DataTransformer schema and updates clients * fix pipeline tests in multinode (does not include automl tests) * fix key leakages by improving cloning of pipeline parameters in AutoML and grids, and by making checksum more consistant * fix TransformerChainTest * fix R CRAN check * add temp changes to understand testMaliciousFreezable failure on Jenkins (can't reproduce on multinode locally) * revert temp log from previous commit + reorder core tests to have water package tested first
valenad1
added a commit
that referenced
this issue
Mar 8, 2024
valenad1
added a commit
that referenced
this issue
Mar 8, 2024
valenad1
added a commit
that referenced
this issue
Mar 11, 2024
* Revert "GH-15857: cleanup legacy TE integration in ModelBuilder and AutoML (#16061)" This reverts commit a8f309b. * Revert "GH-15857: AutoML pipeline support (#16041)" This reverts commit 17fa9ee. * Revert "GH-15856: Grid pipeline support (#16040)" This reverts commit b7ac670. * Revert "GH-15855: core pipeline API (#16039)" This reverts commit c15ea1e.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Sub-issue of #15854
AutoML must provide a mechanism that will activate the use of pipelines during the entire AutoML training, i.e. for base models training, for the HPO steps, and for the Stacked Ensemble models (note that SEs don't need to be encapsulated into Pipeline models, only their base models should be able to apply transformations when generating the metalearner's training frame, as well as when scoring).
** Requirements **
As a first integration example, replace the legacy/hacky integration of Target Encoding to AutoML with the new pipeline: this should allow HPO on some of the TE params during grid search.
The text was updated successfully, but these errors were encountered: