A Julia machine learning framework
Closed issues:
@from_network
does more strange eval
stuff (#703)Clustering.jl
Section (#1000)predict
should work on DataFrameRow
(#1004)reformat
. (#1010)Merged pull requests:
Closed issues:
ProbabilisticSet
type in MLJModelInterface.jl
(#978)Merged pull requests:
bib
, md
, and drawio
from repo stats (#995) (@rikhuijzer)MLJBase compatibility is bumped to 0.21 and MLJModels compatibility is bumped to 0.16. This makes a new simplified method for exporting learning networks available but also introduces some breaking changes:
value
method is no longer exported by MLJ as essentially private (#891)Closed issues:
value
(#891)Merged pull requests:
Closed issues:
reporting_operations
trait (#956)serializable
and restore!
should be "safe" to use any time (#965)serializable
and restore!
(#975)Merged pull requests:
Closed issues:
feature_importances
(#954)Merged pull requests:
Closed issues:
scitype_check_level
(#936)PerformanceEvaluation
(#944)Merged pull requests:
default_scitype_check_level
(#943) (@ablaom)Closed issues:
Save
method documentation (#899)acceleration
to implement parallelism (#934)Merged pull requests:
This release supports changes appearing in the upstream package releases listed below (click on package for detailed release notes).
The principal change, which is breaking, is how model
serialization works. The previous MLJ.save
method still works, but
you can only save to Julia JLS files, and the format is
new and not backwards compatible. A new workflow allows for
serialization using any generic serializer; serialization now plays
nicely with model composition and model wrappers, such as TunedModel
and EnsembleModel
(even with non-Julia atomic models),
and training data will not be inadvertently serialized.
Refer to this manual page details.
The package MLJSerialization has been dropped as a dependency as serialization functionality has moved to MLJBase.
Closed issues:
ScientificTypes
and CategoricalArrays
in native model (#907)Merged pull requests: