Merge pull request #146 from JuliaAI/dev

For a 0.6.11 release
JuliaAI · Sep 10, 2021 · a819b50 · a819b50
2 parents d5d7beb + 4d65e28
commit a819b50
Show file tree

Hide file tree

Showing 5 changed files with 35 additions and 5 deletions.
diff --git a/Project.toml b/Project.toml
@@ -1,7 +1,7 @@
 name = "MLJTuning"
 uuid = "03970b2e-30c4-11ea-3135-d1576263f10f"
 authors = ["Anthony D. Blaom <anthony.blaom@gmail.com>"]
-version = "0.6.10"
+version = "0.6.11"
 
 [deps]
 ComputationalResources = "ed09eef8-17a6-5b46-8889-db040fac31e3"

diff --git a/README.md b/README.md
@@ -4,8 +4,8 @@ Hyperparameter optimization for
 [MLJ](https://github.com/alan-turing-institute/MLJ.jl) machine
 learning models.
 
-[![Build Status](https://github.com/alan-turing-institute/MLJTuning.jl/workflows/CI/badge.svg)](https://github.com/alan-turing-institute/MLJTuning.jl/actions)
-[![Coverage Status](https://coveralls.io/repos/github/alan-turing-institute/MLJTuning.jl/badge.svg?branch=master)](https://coveralls.io/github/alan-turing-institute/MLJTuning.jl?branch=master)
+[![Build Status](https://github.com/JuliaAI/MLJTuning.jl/workflows/CI/badge.svg)](https://github.com/JuliaAI/MLJTuning.jl/actions)
+[![codecov.io](http://codecov.io/github/JuliaAI/MLJTuning.jl/coverage.svg?branch=master)](http://codecov.io/github/JuliaAI/MLJTuning.jl?branch=master)
 
 ### Contents
 
@@ -41,7 +41,7 @@ MLJTuning is a component of the [MLJ
   [MLJModels](https://github.com/alan-turing-institute/MLJModels.jl)
   as a dependency (no ability to search and load registered MLJ
   models). It does however depend on
-  [MLJBase](https://github.com/alan-turing-institute/MLJBase.jl) and,
+  [MLJBase](https://github.com/JuliaAI/MLJBase.jl) and,
   in particular, on the resampling functionality currently residing
   there.
 

diff --git a/src/tuned_models.jl b/src/tuned_models.jl
@@ -691,7 +691,7 @@ function MLJBase.fit(tuned_model::EitherTunedModel{T,M},
     # instantiate resampler (`model` to be replaced with mutated
     # clones during iteration below):
     resampler = Resampler(model=model,
-                          resampling    = tuned_model.resampling,
+                          resampling    = deepcopy(tuned_model.resampling),
                           measure       = tuned_model.measure,
                           weights       = tuned_model.weights,
                           operation     = tuned_model.operation,

diff --git a/test/models/Constant.jl b/test/models/Constant.jl
@@ -2,6 +2,7 @@
 
 const MMI = MLJModelInterface
 export ConstantClassifier, ConstantRegressor,
+       DeterministicConstantRegressor,
        DeterministicConstantClassifier,
        ProbabilisticConstantClassifer
 

diff --git a/test/tuned_models.jl b/test/tuned_models.jl
@@ -265,4 +265,33 @@ end
     @test length(report(mach).history) == 49
 end
 
+@testset_accelerated "Resampling reproducibility" accel begin
+    X, y = make_regression(100, 2)
+    dcr = DeterministicConstantRegressor()
+
+    # Hold out reproducibility
+    homodel = TunedModel(tuning=Explicit(),
+                         models=fill(dcr, 10),
+                         resampling=Holdout(rng=StableRNG(1234)),
+                         acceleration_resampling=accel,
+                         measure=mae)
+    homach = machine(homodel, X, y)
+    fit!(homach, verbosity=0);
+    horep = report(homach)
+    measurements = getproperty.(horep.history, :measurement)
+    @test all(==(measurements[1]), measurements)
+
+    # Cross-validation reproducibility
+    cvmodel = TunedModel(tuning=Explicit(),
+                         models=fill(dcr, 10),
+                         resampling=CV(nfolds=5, rng=StableRNG(1234)),
+                         acceleration_resampling=accel,
+                         measure=mae)
+    cvmach = machine(cvmodel, X, y)
+    fit!(cvmach, verbosity=0);
+    cvrep = report(cvmach)
+    per_folds = getproperty.(cvrep.history, :per_fold)
+    @test all(==(per_folds[1]), per_folds)
+end
+
 true