diff --git a/Project.toml b/Project.toml
index f243bfa..040764f 100644
--- a/Project.toml
+++ b/Project.toml
@@ -6,6 +6,7 @@ version = "0.1.8"
 [deps]
 CategoricalArrays = "324d7699-5711-5eae-9e2f-1d82baa6b597"
 ChainRules = "082447d4-558c-5d27-93f4-14fc19e9eca2"
+ComputationalResources = "ed09eef8-17a6-5b46-8889-db040fac31e3"
 Flux = "587475ba-b771-5e3f-ad9e-33799f191a9c"
 LazyArtifacts = "4af54fe1-eca0-43a8-85a7-787d91b784e3"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
@@ -16,8 +17,11 @@ MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
 MLUtils = "f1d291b0-491e-4a28-83b9-f70985020b54"
 NaturalSort = "c020b1a1-e9b0-503a-9c33-f039bfc54a85"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
+ProgressMeter = "92933f4c-e287-5a05-a399-4b506db050ca"
+Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 Serialization = "9e88b42a-f829-5b0c-bbe9-9e923198166b"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
+Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 
 [compat]
 CategoricalArrays = "0.10"
diff --git a/docs/Manifest.toml b/docs/Manifest.toml
index 37d10a6..2d0743e 100644
--- a/docs/Manifest.toml
+++ b/docs/Manifest.toml
@@ -1,8 +1,8 @@
 # This file is machine-generated - editing it directly is not advised
 
-julia_version = "1.9.2"
+julia_version = "1.9.0"
 manifest_format = "2.0"
-project_hash = "e1685a29d6d370eab88233cc5ac9d849e2f3994f"
+project_hash = "006313002b3cecd4f4a1095930b3c91deb25115f"
 
 [[deps.ANSIColoredPrinters]]
 git-tree-sha1 = "574baf8110975760d391c710b6341da1afa48d8c"
@@ -17,13 +17,14 @@ version = "1.4.1"
 
 [[deps.AbstractFFTs]]
 deps = ["LinearAlgebra"]
-git-tree-sha1 = "cad4c758c0038eea30394b1b671526921ca85b21"
+git-tree-sha1 = "d92ad398961a3ed262d8bf04a1a2b8340f915fef"
 uuid = "621f4979-c628-5d54-868e-fcf4e3e8185c"
-version = "1.4.0"
-weakdeps = ["ChainRulesCore"]
+version = "1.5.0"
+weakdeps = ["ChainRulesCore", "Test"]
 
     [deps.AbstractFFTs.extensions]
     AbstractFFTsChainRulesCoreExt = "ChainRulesCore"
+    AbstractFFTsTestExt = "Test"
 
 [[deps.AbstractPlutoDingetjes]]
 deps = ["Pkg"]
@@ -184,10 +185,10 @@ uuid = "e1450e63-4bb3-523b-b2a4-4ffa8c0fd77d"
 version = "1.2.0"
 
 [[deps.BytePairEncoding]]
-deps = ["StructWalk", "TextEncodeBase", "Unicode"]
-git-tree-sha1 = "40ee2783de5efc5b478e1bb828b750ad8ce1714f"
+deps = ["DoubleArrayTries", "StructWalk", "TextEncodeBase", "Unicode"]
+git-tree-sha1 = "91752c465dfbdd55837a18f9aa9e6d20899658e9"
 uuid = "a4280ba5-8788-555a-8ca8-4a8c3d966a71"
-version = "0.3.1"
+version = "0.3.2"
 
 [[deps.Bzip2_jll]]
 deps = ["Artifacts", "JLLWrappers", "Libdl", "Pkg"]
@@ -315,9 +316,9 @@ version = "0.1.12"
 
 [[deps.Clustering]]
 deps = ["Distances", "LinearAlgebra", "NearestNeighbors", "Printf", "Random", "SparseArrays", "Statistics", "StatsBase"]
-git-tree-sha1 = "42fe66dbc8f1d09a44aa87f18d26926d06a35f84"
+git-tree-sha1 = "b86ac2c5543660d238957dbde5ac04520ae977a7"
 uuid = "aaaa29a8-35af-508c-8bc3-b662a17a0fe5"
-version = "0.15.3"
+version = "0.15.4"
 
 [[deps.CodecZlib]]
 deps = ["TranscodingStreams", "Zlib_jll"]
@@ -327,9 +328,9 @@ version = "0.7.2"
 
 [[deps.ColorSchemes]]
 deps = ["ColorTypes", "ColorVectorSpace", "Colors", "FixedPointNumbers", "PrecompileTools", "Random"]
-git-tree-sha1 = "dd3000d954d483c1aad05fe1eb9e6a715c97013e"
+git-tree-sha1 = "d9a8f86737b665e15a9641ecbac64deef9ce6724"
 uuid = "35d6a980-a343-548e-a6ea-1d62b119f2f4"
-version = "3.22.0"
+version = "3.23.0"
 
 [[deps.ColorTypes]]
 deps = ["FixedPointNumbers", "Random"]
@@ -366,9 +367,9 @@ version = "0.3.0"
 
 [[deps.Compat]]
 deps = ["UUIDs"]
-git-tree-sha1 = "5ce999a19f4ca23ea484e92a1774a61b8ca4cf8e"
+git-tree-sha1 = "e460f044ca8b99be31d35fe54fc33a5c33dd8ed7"
 uuid = "34da2185-b29b-5c13-b0c7-acf172513d20"
-version = "4.8.0"
+version = "4.9.0"
 weakdeps = ["Dates", "LinearAlgebra"]
 
     [deps.Compat.extensions]
@@ -377,7 +378,7 @@ weakdeps = ["Dates", "LinearAlgebra"]
 [[deps.CompilerSupportLibraries_jll]]
 deps = ["Artifacts", "Libdl"]
 uuid = "e66e0078-7015-5450-92f7-15fbd957f2ae"
-version = "1.0.5+0"
+version = "1.0.2+0"
 
 [[deps.CompositionsBase]]
 git-tree-sha1 = "802bb88cd69dfd1509f6670416bd4434015693ad"
@@ -403,7 +404,7 @@ version = "2.2.1"
 
 [[deps.ConformalPrediction]]
 deps = ["CategoricalArrays", "ChainRules", "Flux", "LazyArtifacts", "LinearAlgebra", "MLJBase", "MLJEnsembles", "MLJFlux", "MLJModelInterface", "MLUtils", "NaturalSort", "Plots", "Serialization", "StatsBase"]
-path = ".."
+git-tree-sha1 = "d4ce78a2a13fa4880daf7051a39f888509788d7d"
 uuid = "98bfc277-1877-43dc-819b-a3e38c30242f"
 version = "0.1.8"
 
@@ -470,9 +471,9 @@ version = "1.6.1"
 
 [[deps.DataStructures]]
 deps = ["Compat", "InteractiveUtils", "OrderedCollections"]
-git-tree-sha1 = "cf25ccb972fec4e4817764d01c82386ae94f77b4"
+git-tree-sha1 = "3dbd312d370723b6bb43ba9d02fc36abade4518d"
 uuid = "864edb3b-99cc-5e75-8d2d-829cb0a9cfe8"
-version = "0.18.14"
+version = "0.18.15"
 
 [[deps.DataValueInterfaces]]
 git-tree-sha1 = "bfc1187b79289637fa0ef6d4436ebdfe6905cbd6"
@@ -527,10 +528,10 @@ deps = ["Random", "Serialization", "Sockets"]
 uuid = "8ba89e20-285c-5b6f-9357-94700520ee1b"
 
 [[deps.Distributions]]
-deps = ["FillArrays", "LinearAlgebra", "PDMats", "Printf", "QuadGK", "Random", "SparseArrays", "SpecialFunctions", "Statistics", "StatsAPI", "StatsBase", "StatsFuns", "Test"]
-git-tree-sha1 = "e76a3281de2719d7c81ed62c6ea7057380c87b1d"
+deps = ["FillArrays", "LinearAlgebra", "PDMats", "Printf", "QuadGK", "Random", "SpecialFunctions", "Statistics", "StatsAPI", "StatsBase", "StatsFuns", "Test"]
+git-tree-sha1 = "938fe2981db009f531b6332e31c58e9584a2f9bd"
 uuid = "31c24e10-a181-5473-b8eb-7969acd0382f"
-version = "0.25.98"
+version = "0.25.100"
 
     [deps.Distributions.extensions]
     DistributionsChainRulesCoreExt = "ChainRulesCore"
@@ -571,9 +572,9 @@ version = "0.6.8"
 
 [[deps.DynamicExpressions]]
 deps = ["Compat", "LinearAlgebra", "LoopVectorization", "MacroTools", "PackageExtensionCompat", "PrecompileTools", "Printf", "Random", "Reexport", "TOML"]
-git-tree-sha1 = "a98488649931b24f320dd737a9adf6eb49b5deda"
+git-tree-sha1 = "4f59922b1b80847959c5b1987437d2fcd83c8788"
 uuid = "a40a106e-89c9-4ca8-8020-a735e8728b6b"
-version = "0.11.0"
+version = "0.12.3"
 
     [deps.DynamicExpressions.extensions]
     DynamicExpressionsSymbolicUtilsExt = "SymbolicUtils"
@@ -584,10 +585,10 @@ version = "0.11.0"
     Zygote = "e88e6eb3-aa80-5325-afca-941959d7151f"
 
 [[deps.DynamicQuantities]]
-deps = ["Compat", "LinearAlgebra", "Requires", "SparseArrays", "Tricks"]
-git-tree-sha1 = "c9c36325b0b29b479725d0d8b368081e4b053933"
+deps = ["Compat", "LinearAlgebra", "PackageExtensionCompat", "SparseArrays", "Tricks"]
+git-tree-sha1 = "2a8346673552625beb26680d20b4cfd9285167d0"
 uuid = "06fc5a27-2a28-4c7c-a15d-362465fb6821"
-version = "0.6.2"
+version = "0.6.3"
 weakdeps = ["ScientificTypes", "ScientificTypesBase", "Unitful"]
 
     [deps.DynamicQuantities.extensions]
@@ -608,9 +609,9 @@ version = "0.3.0"
 
 [[deps.EvoTrees]]
 deps = ["BSON", "CUDA", "CategoricalArrays", "Distributions", "MLJModelInterface", "NetworkLayout", "Random", "RecipesBase", "Statistics", "StatsBase", "Tables"]
-git-tree-sha1 = "1b63fdc0acad47c3203398171c138835c1c40d69"
+git-tree-sha1 = "1b418518c0eb1fd1ef0a6d0bfc8051e6abb1232b"
 uuid = "f6006082-12f8-11e9-0c9c-0d5d367ab1e5"
-version = "0.15.0"
+version = "0.15.2"
 
 [[deps.ExceptionUnwrapping]]
 deps = ["Test"]
@@ -625,9 +626,9 @@ uuid = "2e619515-83b5-522b-bb60-26c02a35a201"
 version = "2.5.0+0"
 
 [[deps.ExprTools]]
-git-tree-sha1 = "c1d06d129da9f55715c6c212866f5b1bddc5fa00"
+git-tree-sha1 = "27415f162e6028e81c72b82ef756bf321213b6ec"
 uuid = "e2ba6199-217a-4e67-a87a-7c52f15ade04"
-version = "0.1.9"
+version = "0.1.10"
 
 [[deps.Extents]]
 git-tree-sha1 = "5e1e4c53fa39afe63a7d356e30452249365fba99"
@@ -759,9 +760,9 @@ version = "0.4.2"
 
 [[deps.ForwardDiff]]
 deps = ["CommonSubexpressions", "DiffResults", "DiffRules", "LinearAlgebra", "LogExpFunctions", "NaNMath", "Preferences", "Printf", "Random", "SpecialFunctions"]
-git-tree-sha1 = "00e252f4d706b3d55a8863432e742bf5717b498d"
+git-tree-sha1 = "cf0fe81336da9fb90944683b8c41984b08793dad"
 uuid = "f6369f11-7733-5829-9624-2563aa707210"
-version = "0.10.35"
+version = "0.10.36"
 weakdeps = ["StaticArrays"]
 
     [deps.ForwardDiff.extensions]
@@ -932,9 +933,9 @@ version = "0.3.1"
 
 [[deps.HostCPUFeatures]]
 deps = ["BitTwiddlingConvenienceFunctions", "IfElse", "Libdl", "Static"]
-git-tree-sha1 = "d38bd0d9759e3c6cfa19bdccc314eccf8ce596cc"
+git-tree-sha1 = "eb8fed28f4994600e29beef49744639d985a04b2"
 uuid = "3e5b6fbb-0976-4d2c-9146-d79de83f2fb0"
-version = "0.1.15"
+version = "0.1.16"
 
 [[deps.HuggingFaceApi]]
 deps = ["Dates", "Downloads", "JSON3", "LibGit2", "OhMyArtifacts", "Pkg", "SHA"]
@@ -944,9 +945,9 @@ version = "0.1.0"
 
 [[deps.HypergeometricFunctions]]
 deps = ["DualNumbers", "LinearAlgebra", "OpenLibm_jll", "SpecialFunctions"]
-git-tree-sha1 = "83e95aaab9dc184a6dcd9c4c52aa0dc26cd14a1d"
+git-tree-sha1 = "f218fe3736ddf977e0e772bc9a586b2383da2685"
 uuid = "34004b35-14d8-5ef3-9330-4cdb6864b03a"
-version = "0.3.21"
+version = "0.3.23"
 
 [[deps.Hyperscript]]
 deps = ["Test"]
@@ -1021,9 +1022,9 @@ version = "0.2.17"
 
 [[deps.ImageFiltering]]
 deps = ["CatIndices", "ComputationalResources", "DataStructures", "FFTViews", "FFTW", "ImageBase", "ImageCore", "LinearAlgebra", "OffsetArrays", "PrecompileTools", "Reexport", "SparseArrays", "StaticArrays", "Statistics", "TiledIteration"]
-git-tree-sha1 = "c371a39622dc3b941ffd7c00e6b519d63b3f3f06"
+git-tree-sha1 = "432ae2b430a18c58eb7eca9ef8d0f2db90bc749c"
 uuid = "6a3955dd-da59-5b1f-98d4-e7296123deb5"
-version = "0.7.7"
+version = "0.7.8"
 
 [[deps.ImageIO]]
 deps = ["FileIO", "IndirectArrays", "JpegTurbo", "LazyModules", "Netpbm", "OpenEXR", "PNGFiles", "QOI", "Sixel", "TiffImages", "UUIDs"]
@@ -1120,9 +1121,9 @@ version = "0.1.5"
 
 [[deps.IntelOpenMP_jll]]
 deps = ["Artifacts", "JLLWrappers", "Libdl", "Pkg"]
-git-tree-sha1 = "0cb9352ef2e01574eeebdb102948a58740dcaf83"
+git-tree-sha1 = "ad37c091f7d7daf900963171600d7c1c5c3ede32"
 uuid = "1d5cc7b8-4909-519e-a0f8-d0f5ad9712d0"
-version = "2023.1.0+0"
+version = "2023.2.0+0"
 
 [[deps.InteractiveUtils]]
 deps = ["Markdown"]
@@ -1141,10 +1142,14 @@ uuid = "a98d9a8b-a2ab-59e6-89dd-64a1c18fca59"
 version = "0.14.7"
 
 [[deps.IntervalSets]]
-deps = ["Dates", "Random", "Statistics"]
-git-tree-sha1 = "16c0cc91853084cb5f58a78bd209513900206ce6"
+deps = ["Dates", "Random"]
+git-tree-sha1 = "8e59ea773deee525c99a8018409f64f19fb719e6"
 uuid = "8197267c-284f-5f27-9208-e0e47529a953"
-version = "0.7.4"
+version = "0.7.7"
+weakdeps = ["Statistics"]
+
+    [deps.IntervalSets.extensions]
+    IntervalSetsStatisticsExt = "Statistics"
 
 [[deps.InvertedIndices]]
 git-tree-sha1 = "0dc7b50b8d436461be01300fd8cd45aa0274b038"
@@ -1204,9 +1209,9 @@ version = "0.21.4"
 
 [[deps.JSON3]]
 deps = ["Dates", "Mmap", "Parsers", "PrecompileTools", "StructTypes", "UUIDs"]
-git-tree-sha1 = "5b62d93f2582b09e469b3099d839c2d2ebf5066d"
+git-tree-sha1 = "95220473901735a0f4df9d1ca5b171b568b2daa3"
 uuid = "0f8b85d8-7281-11e9-16c2-39a750bddbf1"
-version = "1.13.1"
+version = "1.13.2"
 
 [[deps.JpegTurbo]]
 deps = ["CEnum", "FileIO", "ImageCore", "JpegTurbo_jll", "TOML"]
@@ -1462,9 +1467,9 @@ weakdeps = ["ChainRulesCore", "ForwardDiff", "SpecialFunctions"]
 
 [[deps.LossFunctions]]
 deps = ["Markdown", "Requires", "Statistics"]
-git-tree-sha1 = "065084a6e63bb30b622b46c613a8f61116787174"
+git-tree-sha1 = "c2b72b61d2e3489b1f9cae3403226b21ec90c943"
 uuid = "30fc2ffe-d236-52d8-8643-a9d8f7c094a7"
-version = "0.10.1"
+version = "0.11.0"
 weakdeps = ["CategoricalArrays"]
 
     [deps.LossFunctions.extensions]
@@ -1483,9 +1488,9 @@ version = "0.1.4"
 
 [[deps.MKL_jll]]
 deps = ["Artifacts", "IntelOpenMP_jll", "JLLWrappers", "LazyArtifacts", "Libdl", "Pkg"]
-git-tree-sha1 = "154d7aaa82d24db6d8f7e4ffcfe596f40bff214b"
+git-tree-sha1 = "eb006abbd7041c28e0d16260e50a24f8f9104913"
 uuid = "856f044c-d86e-5d09-b602-aeab76dc8ba7"
-version = "2023.1.0+0"
+version = "2023.2.0+0"
 
 [[deps.MLDatasets]]
 deps = ["CSV", "Chemfiles", "DataDeps", "DataFrames", "DelimitedFiles", "FileIO", "FixedPointNumbers", "GZip", "Glob", "HDF5", "ImageShow", "JLD2", "JSON3", "LazyModules", "MAT", "MLUtils", "NPZ", "Pickle", "Printf", "Requires", "SparseArrays", "Statistics", "Tables"]
@@ -1501,9 +1506,9 @@ version = "0.19.2"
 
 [[deps.MLJBase]]
 deps = ["CategoricalArrays", "CategoricalDistributions", "ComputationalResources", "Dates", "DelimitedFiles", "Distributed", "Distributions", "InteractiveUtils", "InvertedIndices", "LinearAlgebra", "LossFunctions", "MLJModelInterface", "Missings", "OrderedCollections", "Parameters", "PrettyTables", "ProgressMeter", "Random", "ScientificTypes", "Serialization", "StatisticalTraits", "Statistics", "StatsBase", "Tables"]
-git-tree-sha1 = "9094381ad079dde43c4c74a2f71926232f11cb12"
+git-tree-sha1 = "2c9d6b9c627a80f6e6acbc6193026f455581fd04"
 uuid = "a7f614a8-145f-11e9-1d2a-a57a1082229d"
-version = "0.21.12"
+version = "0.21.13"
 
 [[deps.MLJDecisionTreeInterface]]
 deps = ["CategoricalArrays", "DecisionTree", "MLJModelInterface", "Random", "Tables"]
@@ -1549,9 +1554,9 @@ version = "1.8.0"
 
 [[deps.MLJModels]]
 deps = ["CategoricalArrays", "CategoricalDistributions", "Combinatorics", "Dates", "Distances", "Distributions", "InteractiveUtils", "LinearAlgebra", "MLJModelInterface", "Markdown", "OrderedCollections", "Parameters", "Pkg", "PrettyPrinting", "REPL", "Random", "RelocatableFolders", "ScientificTypes", "StatisticalTraits", "Statistics", "StatsBase", "Tables"]
-git-tree-sha1 = "8f2cf0a7147d370d0de402d43f6de0d3473fcd5e"
+git-tree-sha1 = "2b49f04f70266a2b040eb46ece157c4f5c1b0c13"
 uuid = "d491faf4-2d78-11e9-2867-c94bc002c0b7"
-version = "0.16.9"
+version = "0.16.10"
 
 [[deps.MLJMultivariateStatsInterface]]
 deps = ["CategoricalDistributions", "Distances", "LinearAlgebra", "MLJModelInterface", "MultivariateStats", "StatsBase"]
@@ -1750,9 +1755,9 @@ version = "1.2.0"
 
 [[deps.NeuralAttentionlib]]
 deps = ["Adapt", "CUDA", "ChainRulesCore", "GPUArrays", "GPUArraysCore", "LinearAlgebra", "NNlib", "NNlibCUDA", "Requires", "Static"]
-git-tree-sha1 = "5ee110f3d54e0f29daacc3bdde01b638bf05b9bc"
+git-tree-sha1 = "dab54e810d7d9159c73d3f8b43de9e4b98286517"
 uuid = "12afc1b8-fad6-47e1-9132-84abc478905f"
-version = "0.2.10"
+version = "0.2.11"
 
 [[deps.Observables]]
 git-tree-sha1 = "6862738f9796b3edc1c09d0890afce4eca9e7e93"
@@ -1819,9 +1824,9 @@ version = "1.4.1"
 
 [[deps.OpenSSL_jll]]
 deps = ["Artifacts", "JLLWrappers", "Libdl"]
-git-tree-sha1 = "1aa4b74f80b01c6bc2b89992b861b5f210e665b5"
+git-tree-sha1 = "bbb5c2115d63c2f1451cb70e5ef75e8fe4707019"
 uuid = "458c3c95-2e84-50aa-8efc-19380b2a3a95"
-version = "1.1.21+0"
+version = "1.1.22+0"
 
 [[deps.OpenSpecFun_jll]]
 deps = ["Artifacts", "CompilerSupportLibraries_jll", "JLLWrappers", "Libdl", "Pkg"]
@@ -1870,9 +1875,9 @@ uuid = "f57f5aa1-a3ce-4bc8-8ab9-96f992907883"
 version = "0.4.0"
 
 [[deps.PackageExtensionCompat]]
-git-tree-sha1 = "32f3d52212a8d1c5d589a58851b1f04c97339110"
+git-tree-sha1 = "f9b1e033c2b1205cf30fd119f4e50881316c1923"
 uuid = "65ce6f38-6b18-4e1d-a461-8949797d7930"
-version = "1.0.0"
+version = "1.0.1"
 weakdeps = ["Requires", "TOML"]
 
 [[deps.PaddedViews]]
@@ -1889,9 +1894,9 @@ version = "0.12.3"
 
 [[deps.Parsers]]
 deps = ["Dates", "PrecompileTools", "UUIDs"]
-git-tree-sha1 = "4b2e829ee66d4218e0cef22c0a64ee37cf258c29"
+git-tree-sha1 = "716e24b21538abc91f6205fd1d8363f39b442851"
 uuid = "69de0a69-1ddd-5017-9359-2bf0b02dc9f0"
-version = "2.7.1"
+version = "2.7.2"
 
 [[deps.PartialFunctions]]
 git-tree-sha1 = "b3901ea034cfd8aae57a2fa0dde0b0ea18bad1cb"
@@ -1924,7 +1929,7 @@ version = "0.42.2+0"
 [[deps.Pkg]]
 deps = ["Artifacts", "Dates", "Downloads", "FileWatching", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "REPL", "Random", "SHA", "Serialization", "TOML", "Tar", "UUIDs", "p7zip_jll"]
 uuid = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
-version = "1.9.2"
+version = "1.9.0"
 
 [[deps.PkgVersion]]
 deps = ["Pkg"]
@@ -2459,9 +2464,9 @@ version = "5.10.1+6"
 
 [[deps.SymbolicRegression]]
 deps = ["Compat", "Dates", "Distributed", "DynamicExpressions", "DynamicQuantities", "LineSearches", "LossFunctions", "MLJModelInterface", "MacroTools", "Optim", "PackageExtensionCompat", "Pkg", "PrecompileTools", "Printf", "ProgressBars", "Random", "Reexport", "SpecialFunctions", "StatsBase", "TOML", "Tricks"]
-git-tree-sha1 = "c6eb021000ff7bf49e96d62e6d75070149cf2323"
+git-tree-sha1 = "83d12323cc7cd5b9800cb0c27e5a7b0fdda58438"
 uuid = "8254be44-1295-4e6a-a16d-46603ac705cb"
-version = "0.21.2"
+version = "0.22.2"
 
     [deps.SymbolicRegression.extensions]
     SymbolicRegressionJSON3Ext = "JSON3"
@@ -2567,9 +2572,9 @@ version = "0.4.78"
 
 [[deps.Transformers]]
 deps = ["Base64", "BytePairEncoding", "CUDA", "ChainRulesCore", "DataDeps", "DataStructures", "Dates", "DelimitedFiles", "DoubleArrayTries", "Fetch", "FillArrays", "Flux", "FuncPipelines", "Functors", "HTTP", "HuggingFaceApi", "JSON3", "LightXML", "LinearAlgebra", "Mmap", "NNlib", "NNlibCUDA", "NeuralAttentionlib", "Pickle", "Pkg", "PrimitiveOneHot", "Random", "SHA", "Static", "Statistics", "StringViews", "StructWalk", "TextEncodeBase", "Unicode", "ValSplit", "WordTokenizers", "Zygote"]
-git-tree-sha1 = "35b63543a154cea7e9068f45e67c5fdb7467f2ed"
+git-tree-sha1 = "efa5d0441f3f9e6f3e9e5f64e5dacb736edd13ad"
 uuid = "21ca0261-441d-5938-ace7-c90938fde4d4"
-version = "0.2.6"
+version = "0.2.7"
 
 [[deps.Tricks]]
 git-tree-sha1 = "aadb748be58b492045b4f56166b5188aa63ce549"
@@ -2582,9 +2587,9 @@ uuid = "9d95972d-f1c8-5527-a6e0-b4b365fa01f6"
 version = "1.3.0"
 
 [[deps.URIs]]
-git-tree-sha1 = "074f993b0ca030848b897beff716d93aca60f06a"
+git-tree-sha1 = "b7a5e99f24892b6824a954199a45e9ffcc1c70f0"
 uuid = "5c2747f8-b7ea-4ff2-ba2e-563bfd36b1d4"
-version = "1.4.2"
+version = "1.5.0"
 
 [[deps.UUIDs]]
 deps = ["Random", "SHA"]
@@ -2627,9 +2632,9 @@ version = "3.6.0"
 
 [[deps.Unitful]]
 deps = ["Dates", "LinearAlgebra", "Random"]
-git-tree-sha1 = "c4d2a349259c8eba66a00a540d550f122a3ab228"
+git-tree-sha1 = "64eb17acef1d9734cf09967539818f38093d9b35"
 uuid = "1986cc42-f94f-5a68-af5c-568840ba703d"
-version = "1.15.0"
+version = "1.16.2"
 
     [deps.Unitful.extensions]
     ConstructionBaseUnitfulExt = "ConstructionBase"
@@ -2871,9 +2876,9 @@ version = "1.5.5+0"
 
 [[deps.Zygote]]
 deps = ["AbstractFFTs", "ChainRules", "ChainRulesCore", "DiffRules", "Distributed", "FillArrays", "ForwardDiff", "GPUArrays", "GPUArraysCore", "IRTools", "InteractiveUtils", "LinearAlgebra", "LogExpFunctions", "MacroTools", "NaNMath", "PrecompileTools", "Random", "Requires", "SparseArrays", "SpecialFunctions", "Statistics", "ZygoteRules"]
-git-tree-sha1 = "5be3ddb88fc992a7d8ea96c3f10a49a7e98ebc7b"
+git-tree-sha1 = "e2fe78907130b521619bc88408c859a472c4172b"
 uuid = "e88e6eb3-aa80-5325-afca-941959d7151f"
-version = "0.6.62"
+version = "0.6.63"
 
     [deps.Zygote.extensions]
     ZygoteColorsExt = "Colors"
@@ -2918,7 +2923,7 @@ version = "0.15.1+0"
 [[deps.libblastrampoline_jll]]
 deps = ["Artifacts", "Libdl"]
 uuid = "8e850b90-86db-534c-a0d3-1478176c7d93"
-version = "5.8.0+0"
+version = "5.7.0+0"
 
 [[deps.libfdk_aac_jll]]
 deps = ["Artifacts", "JLLWrappers", "Libdl", "Pkg"]
diff --git a/docs/Project.toml b/docs/Project.toml
index 0db2a5b..c8b9ecf 100644
--- a/docs/Project.toml
+++ b/docs/Project.toml
@@ -23,6 +23,7 @@ MLJLinearModels = "6ee0df7b-362f-4a72-a706-9e79364fb692"
 MLJModelInterface = "e80e1ace-859a-464e-9ed9-23947d8ae3ea"
 MLJMultivariateStatsInterface = "1b6a4a23-ba22-4f51-9698-8599985d3728"
 MLJNaiveBayesInterface = "33e4bacb-b9e2-458e-9a13-5d9a90b235fa"
+MLUtils = "f1d291b0-491e-4a28-83b9-f70985020b54"
 Measures = "442fdcdd-2543-5da2-b0f3-8c86c306513e"
 NaiveBayes = "9bbee03b-0db5-5f46-924f-b5c9c21b8c60"
 NaturalSort = "c020b1a1-e9b0-503a-9c33-f039bfc54a85"
diff --git a/docs/pluto/intro.jl b/docs/pluto/intro.jl
index 996cf91..f664571 100644
--- a/docs/pluto/intro.jl
+++ b/docs/pluto/intro.jl
@@ -7,7 +7,14 @@ using InteractiveUtils
 # This Pluto notebook uses @bind for interactivity. When running this notebook outside of Pluto, the following 'mock version' of @bind gives bound variables a default value (instead of an error).
 macro bind(def, element)
     quote
-        local iv = try Base.loaded_modules[Base.PkgId(Base.UUID("6e696c72-6542-2067-7265-42206c756150"), "AbstractPlutoDingetjes")].Bonds.initial_value catch; b -> missing; end
+        local iv = try
+            Base.loaded_modules[Base.PkgId(
+                Base.UUID("6e696c72-6542-2067-7265-42206c756150"),
+                "AbstractPlutoDingetjes",
+            )].Bonds.initial_value
+        catch
+            b -> missing
+        end
         local el = $(esc(element))
         global $(esc(def)) = Core.applicable(Base.get, el) ? Base.get(el) : iv(el)
         el
@@ -18,19 +25,19 @@ end
 # ╠═╡ show_logs = false
 begin
     using ConformalPrediction
-	using ConformalPrediction: set_size
+    using ConformalPrediction: set_size
     using Distributions
     using EvoTrees: EvoTreeRegressor
-	using LinearAlgebra: norm
+    using LinearAlgebra: norm
     using MLJBase
-	using MLJFlux: NeuralNetworkRegressor
+    using MLJFlux: NeuralNetworkRegressor
     using MLJLinearModels
     using MLJModels
     using NearestNeighborModels: KNNRegressor
     using Plots
     using PlutoUI
-	using Random
-	using StatsBase: quantile
+    using Random
+    using StatsBase: quantile
 end;
 
 # ╔═╡ bc0d7575-dabd-472d-a0ce-db69d242ced8
@@ -46,7 +53,7 @@ Let's start by loading the necessary packages:
 # helper functions
 begin
 
-	# UI stuff
+    # UI stuff
     function multi_slider(vals::Dict; title="")
         return PlutoUI.combine() do Child
             inputs = [
@@ -64,53 +71,57 @@ begin
         end
     end
 
-	# SCP illustration:
-	function illustrate_scp(;cov=0.9, xcord=0.0, ycord=0.0)
-		# Data:
-		n = 250
-		D = 5
-		X, y = make_blobs(n, 2; centers=D, cluster_std=2.0)
-		train, test = partition(eachindex(y), 0.8, shuffle=true)
-		
-		# Model:
-		KNNClassifier = @load KNNClassifier pkg=NearestNeighborModels
-		model = KNNClassifier(;K=10) 
-		
-		# Training:
-		conf_model = conformal_model(model; coverage=cov)
-		mach = machine(conf_model, X, y)
-		fit!(mach, rows=train)
-	
-		# Test set:
-		s_test = predict(mach, rows=test)
-		Xtest = MLJBase.matrix(X)[test,:]
-		i_test = argmin(map(X -> norm(X - [xcord,ycord]), eachrow(Xtest)))
-		x1, x2 = Xtest[i_test,:]
-		y_test = y[test][i_test]
-		
-		# Plotting:
-		p1 = contourf(mach.model, mach.fitresult, X, y, 
-			cbar=false)
-		scatter!([x1], [x2], label="Test point", color=:yellow, ms=10, marker=:star)
-		s = mach.model.scores[:calibration]
-		n_cal = length(s)
-		cov = mach.model.coverage
-		x = 1:n_cal
-		p2 = bar(x, s, title="Cal. scores: 1 - p̂[y]", label="")
-		
-		p3 = bar(x, sort(s), label="", title="(1-α) quantile")
-		q̂ = quantile(s, cov)
-		hline!([q̂], label="q̂", lw=3, ls=:dash, color=2)
-	
-		p̂ = pdf.(predict(mach.model.model, mach.fitresult, [x1 x2])[1], classes(y))
-		p4 = bar(1:D, 1 .- p̂, label="", 
-		  title="Test scores: 1-p̂",ylim=(0.0,1.0),
-		  alpha=map(x -> x < q̂ ? 1.0 : 0.2, 1 .- p̂),
-		  linecolor=map(y -> y == y_test ? :yellow : :black ,1:D),
-		  lw = map(y -> y == y_test ? 5 : 1 ,1:D),)
-		hline!([q̂], label="q̂", lw=3, ls=:dash, color=2)
-		plot(p1,p2,p3,p4,layout=(2,2),size=(550,500),dpi=300)
-	end
+    # SCP illustration:
+    function illustrate_scp(; cov=0.9, xcord=0.0, ycord=0.0)
+        # Data:
+        n = 250
+        D = 5
+        X, y = make_blobs(n, 2; centers=D, cluster_std=2.0)
+        train, test = partition(eachindex(y), 0.8; shuffle=true)
+
+        # Model:
+        KNNClassifier = @load KNNClassifier pkg = NearestNeighborModels
+        model = KNNClassifier(; K=10)
+
+        # Training:
+        conf_model = conformal_model(model; coverage=cov)
+        mach = machine(conf_model, X, y)
+        fit!(mach; rows=train)
+
+        # Test set:
+        s_test = predict(mach; rows=test)
+        Xtest = MLJBase.matrix(X)[test, :]
+        i_test = argmin(map(X -> norm(X - [xcord, ycord]), eachrow(Xtest)))
+        x1, x2 = Xtest[i_test, :]
+        y_test = y[test][i_test]
+
+        # Plotting:
+        p1 = contourf(mach.model, mach.fitresult, X, y; cbar=false)
+        scatter!([x1], [x2]; label="Test point", color=:yellow, ms=10, marker=:star)
+        s = mach.model.scores[:calibration]
+        n_cal = length(s)
+        cov = mach.model.coverage
+        x = 1:n_cal
+        p2 = bar(x, s; title="Cal. scores: 1 - p̂[y]", label="")
+
+        p3 = bar(x, sort(s); label="", title="(1-α) quantile")
+        q̂ = quantile(s, cov)
+        hline!([q̂]; label="q̂", lw=3, ls=:dash, color=2)
+
+        p̂ = pdf.(predict(mach.model.model, mach.fitresult, [x1 x2])[1], classes(y))
+        p4 = bar(
+            1:D,
+            1 .- p̂;
+            label="",
+            title="Test scores: 1-p̂",
+            ylim=(0.0, 1.0),
+            alpha=map(x -> x < q̂ ? 1.0 : 0.2, 1 .- p̂),
+            linecolor=map(y -> y == y_test ? :yellow : :black, 1:D),
+            lw=map(y -> y == y_test ? 5 : 1, 1:D),
+        )
+        hline!([q̂]; label="q̂", lw=3, ls=:dash, color=2)
+        return plot(p1, p2, p3, p4; layout=(2, 2), size=(550, 500), dpi=300)
+    end
 end;
 
 # ╔═╡ be8b2fbb-3b3d-496e-9041-9b8f50872350
@@ -145,20 +156,17 @@ The test point is chosen based on the $x$- and $y$-coordinate that can be specif
 # ╔═╡ ebd25f7a-7cd7-4578-8bd0-1332dd5bc47e
 begin
     illu_dict = Dict(
-        "Coverage" => (0.0:0.05:1.0, 0.9), 
-		"x" => (-20.0:1.0:20.0, 0.0), 
-		"y" => (-20.0:1.0:20.0, 0.0)
+        "Coverage" => (0.0:0.05:1.0, 0.9),
+        "x" => (-20.0:1.0:20.0, 0.0),
+        "y" => (-20.0:1.0:20.0, 0.0),
     )
     @bind illu_specs multi_slider(illu_dict, title="")
 end
 
 # ╔═╡ e1115c04-42c8-4b8a-8845-fc55f63defbf
 begin
-	Random.seed!(123)
-	illustrate_scp(;
-		cov=illu_specs.Coverage,
-		xcord=illu_specs.x,
-		ycord=illu_specs.y)
+    Random.seed!(123)
+    illustrate_scp(; cov=illu_specs.Coverage, xcord=illu_specs.x, ycord=illu_specs.y)
 end
 
 # ╔═╡ b47b9bd5-f4c1-439c-b1a8-ef042aa6adc6
@@ -239,8 +247,8 @@ train, test = partition(eachindex(y), 0.4, 0.4; shuffle=true);
 
 # ╔═╡ 698b1429-f478-45ac-8799-d73f6a0ab869
 begin
-	model_dict = tested_atomic_models[:regression]
-	model_dict[:neural_network] = :(@load NeuralNetworkRegressor pkg = MLJFlux)
+    model_dict = tested_atomic_models[:regression]
+    model_dict[:neural_network] = :(@load NeuralNetworkRegressor pkg = MLJFlux)
 end;
 
 # ╔═╡ a34b8c07-08e0-4a0e-a0f9-8054b41b038b
@@ -359,13 +367,13 @@ In particular, we will call `evaluate!` on our conformal model using `emp_covera
 
 # ╔═╡ d1140af9-608a-4669-9595-aee72ffbaa46
 begin
-	holdout = Holdout(fraction_train=0.8) # pro tip: change this to CV
+    holdout = Holdout(; fraction_train=0.8) # pro tip: change this to CV
     model_evaluation = evaluate!(
-        _mach; 
-		operation=MLJBase.predict, 
-		measure=emp_coverage, 
-		verbosity=0,
-		resampling=holdout,
+        _mach;
+        operation=MLJBase.predict,
+        measure=emp_coverage,
+        verbosity=0,
+        resampling=holdout,
     )
     println("Empirical coverage: $(round(model_evaluation.measurement[1], digits=3))")
 end
@@ -409,19 +417,25 @@ Below you can test out other available regression methods. While you may not be
 
 # ╔═╡ 45212d6a-2a09-4a6e-aa39-16cc659d1e18
 begin
-	# Predictions:
-	new_conf_model = conformal_model(model; method=conf_model_name, coverage=new_cov)
-	new_mach = machine(new_conf_model, X, y)
-	MLJBase.fit!(new_mach; rows=train, verbosity=0)
-    p1 = plot(new_mach.model, new_mach.fitresult, Xtest, ytest; 
-		zoom=0, observed_lab="Test points", title="Predictions")
-    plot!(p1, xrange, @.(f(xrange)); 
-		lw=2, ls=:dash, colour=:black, label="Ground truth")
-
-	# Interval width:
-	p2 = bar(new_mach.model, new_mach.fitresult, Xtest, title="Interval Width")
-	
-	plot(p1, p2, size=(1000,400))
+    # Predictions:
+    new_conf_model = conformal_model(model; method=conf_model_name, coverage=new_cov)
+    new_mach = machine(new_conf_model, X, y)
+    MLJBase.fit!(new_mach; rows=train, verbosity=0)
+    p1 = plot(
+        new_mach.model,
+        new_mach.fitresult,
+        Xtest,
+        ytest;
+        zoom=0,
+        observed_lab="Test points",
+        title="Predictions",
+    )
+    plot!(p1, xrange, @.(f(xrange)); lw=2, ls=:dash, colour=:black, label="Ground truth")
+
+    # Interval width:
+    p2 = bar(new_mach.model, new_mach.fitresult, Xtest; title="Interval Width")
+
+    plot(p1, p2; size=(1000, 400))
 end
 
 # ╔═╡ 74444c01-1a0a-47a7-9b14-749946614f07
diff --git a/docs/pluto/jcon2023.jl b/docs/pluto/jcon2023.jl
index 0dd781c..6175335 100644
--- a/docs/pluto/jcon2023.jl
+++ b/docs/pluto/jcon2023.jl
@@ -10,7 +10,7 @@ begin
     using Distributions
     using EvoTrees: EvoTreeRegressor
     using MLJBase
-	using MLJFlux: NeuralNetworkRegressor
+    using MLJFlux: NeuralNetworkRegressor
     using MLJLinearModels
     using MLJModels
     using NearestNeighborModels: KNNRegressor
diff --git a/docs/pluto/understanding_coverage.jl b/docs/pluto/understanding_coverage.jl
index 9132887..5c38505 100644
--- a/docs/pluto/understanding_coverage.jl
+++ b/docs/pluto/understanding_coverage.jl
@@ -7,7 +7,14 @@ using InteractiveUtils
 # This Pluto notebook uses @bind for interactivity. When running this notebook outside of Pluto, the following 'mock version' of @bind gives bound variables a default value (instead of an error).
 macro bind(def, element)
     quote
-        local iv = try Base.loaded_modules[Base.PkgId(Base.UUID("6e696c72-6542-2067-7265-42206c756150"), "AbstractPlutoDingetjes")].Bonds.initial_value catch; b -> missing; end
+        local iv = try
+            Base.loaded_modules[Base.PkgId(
+                Base.UUID("6e696c72-6542-2067-7265-42206c756150"),
+                "AbstractPlutoDingetjes",
+            )].Bonds.initial_value
+        catch
+            b -> missing
+        end
         local el = $(esc(element))
         global $(esc(def)) = Core.applicable(Base.get, el) ? Base.get(el) : iv(el)
         el
@@ -22,7 +29,7 @@ begin
     using EvoTrees: EvoTreeRegressor
     using LightGBM.MLJInterface: LGBMRegressor
     using MLJBase
-	using MLJLinearModels
+    using MLJLinearModels
     using MLJModels
     using NearestNeighborModels: KNNRegressor
     using Plots
diff --git a/docs/src/tutorials/training.qmd b/docs/src/tutorials/training.qmd
new file mode 100644
index 0000000..209c89e
--- /dev/null
+++ b/docs/src/tutorials/training.qmd
@@ -0,0 +1,73 @@
+# ConformalTraining
+
+```@meta
+CurrentModule = ConformalPrediction
+```
+
+```{julia}
+#| echo: false
+using Pkg; Pkg.activate("docs")
+using Plots
+theme(:wong)
+using Random
+Random.seed!(2022)
+www_path = "docs/src/www"   # output path for files don't get automatically saved in auto-generated path (e.g. GIFs)
+```
+
+
+```{julia}
+using MLJ
+using Random
+Random.seed!(123)
+
+# Data:
+X, y = make_blobs(500, centers=4, cluster_std=1.0)
+X = MLJ.table(Float32.(MLJ.matrix(X)))
+train, test = partition(eachindex(y), 0.8, shuffle=true)
+```
+
+
+```{julia}
+using Flux
+using MLJFlux
+using ConformalPrediction
+using ConformalPrediction.ConformalTraining: ConformalNNClassifier
+
+# Model:
+builder = MLJFlux.MLP(hidden=(32, 32, 32,), σ=Flux.relu)
+# clf = ConformalNNClassifier(epochs=250, builder=builder, batch_size=50)
+clf = NeuralNetworkClassifier(epochs=250, builder=builder, batch_size=50)
+```
+
+
+```{julia}
+using ConformalPrediction
+
+conf_model = conformal_model(clf; method=:simple_inductive)
+mach = machine(conf_model, X, y)
+fit!(mach, rows=train)
+```
+
+```{julia}
+#| output: true
+
+using Plots
+p_proba = contourf(mach.model, mach.fitresult, X, y)
+p_set_size = contourf(mach.model, mach.fitresult, X, y; plot_set_size=true)
+p_smooth = contourf(mach.model, mach.fitresult, X, y; plot_set_loss=true)
+plot(p_proba, p_set_size, p_smooth, layout=(1,3), size=(1200,250))
+```
+
+```{julia}
+#| output: true
+
+_eval = evaluate!(
+    mach,
+    operation=predict,
+    measure=[emp_coverage, ssc, ineff]
+)
+
+println("Empirical coverage: $(round(_eval.measurement[1], digits=3))")
+println("SSC: $(round(_eval.measurement[2], digits=3))")
+println("Inefficiency: $(round(_eval.measurement[3], digits=3))")
+```
\ No newline at end of file
diff --git a/src/ConformalPrediction.jl b/src/ConformalPrediction.jl
index 4d6abec..a4c4c00 100644
--- a/src/ConformalPrediction.jl
+++ b/src/ConformalPrediction.jl
@@ -10,7 +10,7 @@ export soft_assignment
 
 # Evaluation:
 include("evaluation/evaluation.jl")
-export emp_coverage, size_stratified_coverage, ssc
+export emp_coverage, size_stratified_coverage, ssc, ineff
 
 # Artifacts:
 include("artifacts/core.jl")
diff --git a/src/conformal_models/ConformalTraining/ConformalTraining.jl b/src/conformal_models/ConformalTraining/ConformalTraining.jl
new file mode 100644
index 0000000..405082e
--- /dev/null
+++ b/src/conformal_models/ConformalTraining/ConformalTraining.jl
@@ -0,0 +1,15 @@
+module ConformalTraining
+
+using ConformalPrediction
+using Flux
+using MLJFlux
+
+const default_builder = MLJFlux.MLP(; hidden=(32, 32, 32), σ=Flux.relu)
+
+include("losses.jl")
+include("inductive_classification.jl")
+include("classifier.jl")
+include("regressor.jl")
+include("training.jl")
+
+end
diff --git a/src/conformal_models/ConformalTraining/classifier.jl b/src/conformal_models/ConformalTraining/classifier.jl
new file mode 100644
index 0000000..2e07ccd
--- /dev/null
+++ b/src/conformal_models/ConformalTraining/classifier.jl
@@ -0,0 +1,91 @@
+using ComputationalResources
+using Flux
+using MLJFlux
+import MLJModelInterface as MMI
+using ProgressMeter
+using Random
+using Tables
+
+"The `ConformalNNClassifier` struct is a wrapper for a `ConformalModel` that can be used with MLJFlux.jl."
+mutable struct ConformalNNClassifier{B,F,O,L} <: MLJFlux.MLJFluxProbabilistic
+    builder::B
+    finaliser::F
+    optimiser::O   # mutable struct from Flux/src/optimise/optimisers.jl
+    loss::L        # can be called as in `loss(yhat, y)`
+    epochs::Int    # number of epochs
+    batch_size::Int  # size of a batch
+    lambda::Float64  # regularization strength
+    alpha::Float64   # regularizaton mix (0 for all l2, 1 for all l1)
+    rng::Union{AbstractRNG,Int64}
+    optimiser_changes_trigger_retraining::Bool
+    acceleration::AbstractResource  # eg, `CPU1()` or `CUDALibs()`
+end
+
+function ConformalNNClassifier(;
+    builder::B=default_builder,
+    finaliser::F=Flux.softmax,
+    optimiser::O=Flux.Optimise.Adam(),
+    loss::L=Flux.crossentropy,
+    epochs::Int=100,
+    batch_size::Int=100,
+    lambda::Float64=0.0,
+    alpha::Float64=0.0,
+    rng::Union{AbstractRNG,Int64}=Random.GLOBAL_RNG,
+    optimiser_changes_trigger_retraining::Bool=false,
+    acceleration::AbstractResource=CPU1(),
+) where {B,F,O,L}
+
+    # Initialise the MLJFlux wrapper:
+    mod = ConformalNNClassifier(
+        builder,
+        finaliser,
+        optimiser,
+        loss,
+        epochs,
+        batch_size,
+        lambda,
+        alpha,
+        rng,
+        optimiser_changes_trigger_retraining,
+        acceleration,
+    )
+
+    return mod
+end
+
+# if `b` is a builder, then `b(model, rng, shape...)` is called to make a
+# new chain, where `shape` is the return value of this method:
+function MLJFlux.shape(model::ConformalNNClassifier, X, y)
+    levels = MMI.classes(y[1])
+    n_output = length(levels)
+    n_input = Tables.schema(X).names |> length
+    return (n_input, n_output)
+end
+
+# builds the end-to-end Flux chain needed, given the `model` and `shape`:
+function MLJFlux.build(model::ConformalNNClassifier, rng, shape)
+
+    # Chain:
+    chain = Flux.Chain(MLJFlux.build(model.builder, rng, shape...), model.finaliser)
+
+    return chain
+end
+
+# returns the model `fitresult` (see "Adding Models for General Use"
+# section of the MLJ manual) which must always have the form `(chain,
+# metadata)`, where `metadata` is anything extra needed by `predict`:
+MLJFlux.fitresult(model::ConformalNNClassifier, chain, y) = (chain, MMI.classes(y[1]))
+
+function MMI.predict(model::ConformalNNClassifier, fitresult, Xnew)
+    chain, levels = fitresult
+    X = MLJFlux.reformat(Xnew)
+    probs = vcat([chain(MLJFlux.tomat(X[:, i]))' for i in 1:size(X, 2)]...)
+    return MMI.UnivariateFinite(levels, probs)
+end
+
+MMI.metadata_model(
+    ConformalNNClassifier;
+    input=Union{AbstractArray,MMI.Table(MMI.Continuous)},
+    target=AbstractVector{<:MMI.Finite},
+    path="MLJFlux.ConformalNNClassifier",
+)
diff --git a/src/conformal_models/training/inductive_classification.jl b/src/conformal_models/ConformalTraining/inductive_classification.jl
similarity index 80%
rename from src/conformal_models/training/inductive_classification.jl
rename to src/conformal_models/ConformalTraining/inductive_classification.jl
index 990cf78..8a76eee 100644
--- a/src/conformal_models/training/inductive_classification.jl
+++ b/src/conformal_models/ConformalTraining/inductive_classification.jl
@@ -1,21 +1,22 @@
+using CategoricalArrays
+using ConformalPrediction: SimpleInductiveClassifier, AdaptiveInductiveClassifier
 using MLJEnsembles: EitherEnsembleModel
 using MLJFlux: MLJFluxModel, reformat
 using MLUtils
 
 """
-    score(conf_model::InductiveModel, model::MLJFluxModel, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
+    ConformalPrediction.score(conf_model::InductiveModel, model::MLJFluxModel, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
 
 Overloads the `score` function for the `MLJFluxModel` type.
 """
-function score(
+function ConformalPrediction.score(
     conf_model::SimpleInductiveClassifier,
     ::Type{<:MLJFluxModel},
     fitresult,
     X,
     y::Union{Nothing,AbstractArray}=nothing,
 )
-    X = reformat(X)
-    X = typeof(X) <: AbstractArray ? X : permutedims(matrix(X))
+    X = permutedims(matrix(X))
     probas = permutedims(fitresult[1](X))
     scores = @.(conf_model.heuristic(probas))
     if isnothing(y)
@@ -27,19 +28,18 @@ function score(
 end
 
 """
-    score(conf_model::SimpleInductiveClassifier, ::Type{<:EitherEnsembleModel{<:MLJFluxModel}}, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
+    ConformalPrediction.score(conf_model::SimpleInductiveClassifier, ::Type{<:EitherEnsembleModel{<:MLJFluxModel}}, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
 
 Overloads the `score` function for ensembles of `MLJFluxModel` types.
 """
-function score(
+function ConformalPrediction.score(
     conf_model::SimpleInductiveClassifier,
     ::Type{<:EitherEnsembleModel{<:MLJFluxModel}},
     fitresult,
     X,
     y::Union{Nothing,AbstractArray}=nothing,
 )
-    X = reformat(X)
-    X = typeof(X) <: AbstractArray ? X : permutedims(matrix(X))
+    X = permutedims(matrix(X))
     _chains = map(res -> res[1], fitresult.ensemble)
     probas =
         MLUtils.stack(map(chain -> chain(X), _chains)) |>
@@ -56,7 +56,7 @@ function score(
 end
 
 """
-    score(conf_model::AdaptiveInductiveClassifier, ::Type{<:MLJFluxModel}, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
+    ConformalPrediction.score(conf_model::AdaptiveInductiveClassifier, ::Type{<:MLJFluxModel}, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
 
 Overloads the `score` function for the `MLJFluxModel` type.
 """
@@ -87,7 +87,7 @@ function score(
 end
 
 """
-    score(conf_model::AdaptiveInductiveClassifier, ::Type{<:EitherEnsembleModel{<:MLJFluxModel}}, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
+    ConformalPrediction.score(conf_model::AdaptiveInductiveClassifier, ::Type{<:EitherEnsembleModel{<:MLJFluxModel}}, fitresult, X, y::Union{Nothing,AbstractArray}=nothing)
 
 Overloads the `score` function for ensembles of `MLJFluxModel` types.
 """
diff --git a/src/conformal_models/ConformalTraining/inductive_regression.jl b/src/conformal_models/ConformalTraining/inductive_regression.jl
new file mode 100644
index 0000000..455af5f
--- /dev/null
+++ b/src/conformal_models/ConformalTraining/inductive_regression.jl
@@ -0,0 +1,42 @@
+using CategoricalArrays
+using ConformalPrediction: SimpleInductiveRegressor
+using MLJEnsembles: EitherEnsembleModel
+using MLJFlux: MLJFluxModel
+using MLUtils
+
+"""
+    ConformalPrediction.score(conf_model::SimpleInductiveRegressor, model::MLJFluxModel, fitresult, X, y)
+
+Overloads the `score` function for the `MLJFluxModel` type.
+"""
+function ConformalPrediction.score(
+    conf_model::SimpleInductiveRegressor, ::Type{<:MLJFluxModel}, fitresult, X, y
+)
+    X = permutedims(matrix(X))
+    ŷ = permutedims(fitresult[1](X))
+    scores = @.(conf_model.heuristic(y, ŷ))
+    return scores
+end
+
+"""
+    ConformalPrediction.score(conf_model::SimpleInductiveRegressor, ::Type{<:EitherEnsembleModel{<:MLJFluxModel}}, fitresult, X, y)
+
+Overloads the `score` function for ensembles of `MLJFluxModel` types.
+"""
+function ConformalPrediction.score(
+    conf_model::SimpleInductiveRegressor,
+    ::Type{<:EitherEnsembleModel{<:MLJFluxModel}},
+    fitresult,
+    X,
+    y,
+)
+    X = permutedims(matrix(X))
+    _chains = map(res -> res[1], fitresult.ensemble)
+    ŷ =
+        MLUtils.stack(map(chain -> chain(X), _chains)) |>
+        y ->
+            mean(y; dims=ndims(y)) |>
+            y -> MLUtils.unstack(y; dims=ndims(y))[1] |> y -> permutedims(y)
+    scores = @.(conf_model.heuristic(y, ŷ))
+    return scores
+end
diff --git a/src/conformal_models/training/losses.jl b/src/conformal_models/ConformalTraining/losses.jl
similarity index 91%
rename from src/conformal_models/training/losses.jl
rename to src/conformal_models/ConformalTraining/losses.jl
index 6a76271..8fc6713 100644
--- a/src/conformal_models/training/losses.jl
+++ b/src/conformal_models/ConformalTraining/losses.jl
@@ -1,6 +1,8 @@
+using ConformalPrediction: ConformalProbabilisticSet
 using Flux
 using LinearAlgebra
 using MLJBase
+using StatsBase
 
 """
     soft_assignment(conf_model::ConformalProbabilisticSet; temp::Real=0.5)
@@ -27,8 +29,8 @@ function soft_assignment(
 )
     temp = isnothing(temp) ? 0.5 : temp
     v = sort(conf_model.scores[:calibration])
-    q̂ = qplus(v, conf_model.coverage; sorted=true)
-    scores = score(conf_model, fitresult, X)
+    q̂ = StatsBase.quantile(v, conf_model.coverage; sorted=true)
+    scores = ConformalPrediction.score(conf_model, fitresult, X)
     return @.(σ((q̂ - scores) / temp))
 end
 
@@ -53,7 +55,7 @@ function smooth_size_loss(
     temp::Union{Nothing,Real}=nothing,
     κ::Real=1.0,
 )
-    temp = isnothing(temp) ? 0.5 : temp
+    temp = isnothing(temp) ? 0.1 : temp
     C = soft_assignment(conf_model, fitresult, X; temp=temp)
     is_empty_set = all(
         x -> x .== 0, soft_assignment(conf_model, fitresult, X; temp=0.0); dims=2
@@ -70,14 +72,6 @@ function smooth_size_loss(
         Ω = [Ω..., ω]
     end
     Ω = permutedims(permutedims(Ω))
-    # Ω = map(sum(C; dims=2), is_empty_set) do x, is_empty
-    #     if is_empty #&& κ > 0
-    #         ω = maximum([x - κ, full_set_size - κ])
-    #     else 
-    #         ω = maximum([0, x - κ])
-    #     end
-    #     return ω
-    # end
     return Ω
 end
 
diff --git a/src/conformal_models/ConformalTraining/regressor.jl b/src/conformal_models/ConformalTraining/regressor.jl
new file mode 100644
index 0000000..3522750
--- /dev/null
+++ b/src/conformal_models/ConformalTraining/regressor.jl
@@ -0,0 +1,82 @@
+using ComputationalResources
+using Flux
+using MLJFlux
+import MLJModelInterface as MMI
+using ProgressMeter
+using Random
+using Tables
+
+"The `ConformalNNRegressor` struct is a wrapper for a `ConformalModel` that can be used with MLJFlux.jl."
+mutable struct ConformalNNRegressor{B,O,L} <: MLJFlux.MLJFluxDeterministic
+    builder::B
+    optimiser::O  # mutable struct from Flux/src/optimise/optimisers.jl
+    loss::L       # can be called as in `loss(yhat, y)`
+    epochs::Int   # number of epochs
+    batch_size::Int # size of a batch
+    lambda::Float64 # regularization strength
+    alpha::Float64  # regularizaton mix (0 for all l2, 1 for all l1)
+    rng::Union{AbstractRNG,Integer}
+    optimiser_changes_trigger_retraining::Bool
+    acceleration::AbstractResource  # eg, `CPU1()` or `CUDALibs()`
+end
+
+function ConformalNNRegressor(;
+    builder::B=default_builder,
+    optimiser::O=Flux.Optimise.Adam(),
+    loss::L=Flux.mse,
+    epochs::Int=100,
+    batch_size::Int=100,
+    lambda::Float64=0.0,
+    alpha::Float64=0.0,
+    rng::Union{AbstractRNG,Int64}=Random.GLOBAL_RNG,
+    optimiser_changes_trigger_retraining::Bool=false,
+    acceleration::AbstractResource=CPU1(),
+) where {B,O,L}
+
+    # Initialise the MLJFlux wrapper:
+    mod = ConformalNNRegressor(
+        builder,
+        optimiser,
+        loss,
+        epochs,
+        batch_size,
+        lambda,
+        alpha,
+        rng,
+        optimiser_changes_trigger_retraining,
+        acceleration,
+    )
+
+    return mod
+end
+
+"""
+    shape(model::NeuralNetworkRegressor, X, y)
+
+A private method that returns the shape of the input and output of the model for given data `X` and `y`.
+"""
+function MLJFlux.shape(model::ConformalNNRegressor, X, y)
+    X = X isa Matrix ? Tables.table(X) : X
+    n_input = Tables.schema(X).names |> length
+    n_ouput = 1
+    return (n_input, 1)
+end
+
+function MLJFlux.build(model::ConformalNNRegressor, rng, shape)
+    return MLJFlux.build(model.builder, rng, shape...)
+end
+
+MLJFlux.fitresult(model::ConformalNNRegressor, chain, y) = (chain, nothing)
+
+function MMI.predict(model::ConformalNNRegressor, fitresult, Xnew)
+    chain = fitresult[1]
+    Xnew_ = MLJFlux.reformat(Xnew)
+    return [chain(values.(MLJFlux.tomat(Xnew_[:, i])))[1] for i in 1:size(Xnew_, 2)]
+end
+
+MMI.metadata_model(
+    ConformalNNRegressor;
+    input=Union{AbstractMatrix{Continuous},MMI.Table(MMI.Continuous)},
+    target=AbstractVector{<:MMI.Continuous},
+    path="MLJFlux.ConformalNNRegressor",
+)
diff --git a/src/conformal_models/ConformalTraining/training.jl b/src/conformal_models/ConformalTraining/training.jl
new file mode 100644
index 0000000..ae8a82e
--- /dev/null
+++ b/src/conformal_models/ConformalTraining/training.jl
@@ -0,0 +1,55 @@
+const ConformalNN = Union{ConformalNNClassifier,ConformalNNRegressor}
+
+@doc raw"""
+    MLJFlux.train!(model::ConformalNN, penalty, chain, optimiser, X, y)
+
+Implements the conformal traning procedure for the `ConformalNN` type.
+"""
+function MLJFlux.train!(model::ConformalNN, penalty, chain, optimiser, X, y)
+
+    # Setup:
+    loss = model.loss
+    n_batches = length(y)
+    training_loss = zero(Float32)
+    size_loss = zero(Float32)
+    fitresult = (chain, nothing)
+
+    # Training loop:
+    for i in 1:n_batches
+        parameters = Flux.params(chain)
+
+        # Data Splitting:
+        X_batch, y_batch = X[i], y[i]
+        conf_model = ConformalPrediction.conformal_model(
+            model; method=:simple_inductive, coverage=0.95
+        )
+        calibration, pred = partition(
+            1:size(y_batch, 2), conf_model.train_ratio; shuffle=true
+        )
+        Xcal = X_batch[:, calibration]
+        ycal = y_batch[:, calibration]
+        Xcal, ycal = MMI.reformat(conf_model.model, Xcal, ycal)
+        Xpred = X_batch[:, pred]
+        ypred = y_batch[:, pred]
+        Xpred, ypred = MMI.reformat(conf_model.model, Xpred, ypred)
+
+        # On-the-fly calibration:
+        cal_scores, scores = ConformalPrediction.score(
+            conf_model, fitresult, Xcal', categorical(Flux.onecold(ycal))
+        )
+        conf_model.scores = Dict(:calibration => cal_scores, :all => scores)
+
+        gs = Flux.gradient(parameters) do
+            Ω = smooth_size_loss(conf_model, fitresult, Xpred')
+            yhat = chain(X_batch)
+            batch_loss = loss(yhat, y_batch) + penalty(parameters) / n_batches
+            batch_loss += 0.5 * sum(Ω) / length(Ω)                            # add size loss
+            training_loss += batch_loss
+            size_loss += sum(Ω) / length(Ω)
+            return batch_loss
+        end
+        Flux.update!(optimiser, parameters, gs)
+    end
+
+    return training_loss / n_batches
+end
diff --git a/src/conformal_models/conformal_models.jl b/src/conformal_models/conformal_models.jl
index ce5cd19..7c1068d 100644
--- a/src/conformal_models/conformal_models.jl
+++ b/src/conformal_models/conformal_models.jl
@@ -57,7 +57,8 @@ include("inductive_classification.jl")
 include("transductive_classification.jl")
 
 # Training:
-include("training/training.jl")
+include("ConformalTraining/ConformalTraining.jl")
+using .ConformalTraining
 
 # Type unions:
 const InductiveModel = Union{
diff --git a/src/conformal_models/plotting.jl b/src/conformal_models/plotting.jl
index 3461499..5e4f55d 100644
--- a/src/conformal_models/plotting.jl
+++ b/src/conformal_models/plotting.jl
@@ -159,7 +159,6 @@ function Plots.contourf(
             kwargs...,
         )
     else
-        clim = @isdefined(clim) ? clim : (0, 1)
         plt = contourf(
             x1range,
             x2range,
@@ -167,7 +166,6 @@ function Plots.contourf(
             title=title,
             xlims=xlims,
             ylims=ylims,
-            clim=clim,
             c=cgrad(:blues),
             linewidth=0,
             kwargs...,
diff --git a/src/conformal_models/training/training.jl b/src/conformal_models/training/training.jl
deleted file mode 100644
index 8fd1da3..0000000
--- a/src/conformal_models/training/training.jl
+++ /dev/null
@@ -1,2 +0,0 @@
-include("losses.jl")
-include("inductive_classification.jl")
diff --git a/src/evaluation/measures.jl b/src/evaluation/measures.jl
index 09ba9f4..f96a4e3 100644
--- a/src/evaluation/measures.jl
+++ b/src/evaluation/measures.jl
@@ -38,3 +38,14 @@ function size_stratified_coverage(ŷ, y)
 
     return C̄
 end
+
+"""
+    ineff(ŷ)
+
+Computes the inefficiency (average set size) for conformal predictions `ŷ`.
+"""
+function ineff(ŷ, y=missing)
+    R = length(ŷ)
+    ineff = sum(set_size.(ŷ)) / R
+    return ineff
+end