Releases: JuliaAI/MLJXGBoostInterface.jl
Releases · JuliaAI/MLJXGBoostInterface.jl
v0.3.0
MLJXGBoostInterface v0.3.0
- Reorganize code for better maintenance and make adjustments for major overhaul of XGBoost.jl, 2.0.
- Fix
save
method to create persistent representations of learned parameters that avoid file I/O.
Merged pull requests:
- elminate code duplication to make this much easier to maintain (#21) (@ExpandingMan)
v0.2.1
MLJXGBoostInterface v0.2.1
Closed issues:
- XGBoostRegressor failing on Boston data set (#17)
Merged pull requests:
- Fix links in README.md (#16) (@tylerjthomas9)
- fix issue #17 (#18) (@OkonSamuel)
v0.2.0
MLJXGBoostInterface v0.2.0
-
(breaking) Re-implement the
save
andrestore
methods supporting serialization, to address MLJ API changes at JuliaAI/MLJBase.jl#733. These changes are not backwards compatible: to restore models serialised using MLJXGBoostInterface < 0.2, you will need a version of MLJXGBoostInterface < 0.2 loaded. -
Bump minimum Julia requirement to 1.6
Merged pull requests:
v0.1.5
v0.1.4
MLJXGBoostInterface v0.1.4
Closed issues:
- XGBoostClassifier is broken for binary classifiation (#7)
Merged pull requests:
- Fix binary classification bug #7 (#9) (@AttilaKovacs)
v0.1.3
MLJXGBoostInterface v0.1.3
Merged pull requests: