Skip to content

Latest commit

 

History

History
56 lines (41 loc) · 1.56 KB

changelog.rst

File metadata and controls

56 lines (41 loc) · 1.56 KB
Raw

Changelog

Version 1.5.0

  • Improved scikit-learn compatibility
  • Fixed bug where fitted estimators were unpicklable

Version 1.4.1

  • Fixed so GroupLasso objects can be initialised without specifying the groups
  • Fixed bug with viewing the selected groups
  • Changed to ConvergenceWarning instead of RuntimeWarning for when FISTA doesn't converge
  • Fixed typo in warning message
  • Fixed bug with list input for the groups
  • Added matplotlib to requirements.txt
  • Changed how groups are parsed, should hopefully fix some unreported bugs

Version 1.4.0

  • Fixed bug with how LogisticGroupLasso applied intercept in some methods
  • Implemented backtracking line search for Lipschitz
  • Added centering-based preconditioning for dense design matrices

Version 1.2.0

  • Merged logistic group lasso and multinomial group lasso to one class.
  • Added support for sparse matrices.
  • Added utility for extracting groups from a sklearn OneHotEncoder.
  • Added option to control group-wise scaling of regularisation.

Version 1.2.2

  • Fixed bug where the old_regularisation and supress_warning arguments could only be set by the base group lasso class, but none of the actual models.

Version 1.2.1

  • Fixed bug where the regularisation was multiplied with the Lipschitz coefficient of the loss gradient.

Version 1.1.1

  • Added warm restarts

Version 1.0.0

  • Started changelog
  • Working group LASSO regularised least squares, logistic and multinomial regression.