You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Auto-sklearn was built for the AutoML challenges where there were public and private test sets by the names "validation set" and "test set". Since these challenges we have substantially refactored Auto-sklearn and no longer allow the user to pass in the public test set aka "validation set". This leads to confusion as "validation set" is really a misnomer, as inside a machine learning system it usually denotes the dataset on which a model is selected. Therefore we should drop all references that still exist to this old concept of a "validation set", as already being done by @eddiebergman in #1434, which should only be in the evaluation submodule`.
The text was updated successfully, but these errors were encountered:
Auto-sklearn was built for the AutoML challenges where there were public and private test sets by the names "validation set" and "test set". Since these challenges we have substantially refactored Auto-sklearn and no longer allow the user to pass in the public test set aka "validation set". This leads to confusion as "validation set" is really a misnomer, as inside a machine learning system it usually denotes the dataset on which a model is selected. Therefore we should drop all references that still exist to this old concept of a "validation set", as already being done by @eddiebergman in #1434, which should only be in the
evaluation
submodule`.The text was updated successfully, but these errors were encountered: