Merge pull request #123 from juliohm/crossentropy

Rename CrossentropyLoss to CrossEntropyLoss

src/LossFunctions.jl

@@ -63,7 +63,7 @@ export
 
     PoissonLoss,
     LogitProbLoss,
-    CrossentropyLoss,
+    CrossEntropyLoss,
     ZeroOneLoss,
 
     OrdinalMarginLoss,

src/deprecated.jl

@@ -5,6 +5,7 @@
 @deprecate scaled(loss, ::Type{Val{K}}) where {K}  scaled(loss, Val(K))
 @deprecate weightedloss(loss::Loss, ::Type{Val{W}}) where {W}  weightedloss(loss, Val(W))
 @deprecate WeightedBinaryLoss(loss, ::Type{Val{W}}) where {W}  WeightedBinaryLoss(loss, Val(W))
+@deprecate CrossentropyLoss CrossEntropyLoss
 
 Base.@deprecate_binding AverageMode AggregateMode
 export AvgMode

src/supervised/other.jl

@@ -27,16 +27,16 @@ isstronglyconvex(::PoissonLoss) = false
 # ===============================================================
 
 @doc doc"""
-    CrossentropyLoss <: SupervisedLoss
+    CrossEntropyLoss <: SupervisedLoss
 
 Cross-entropy loss also known as log loss and logistic loss is defined as:
 
 ``L(target, output) = - target*ln(output) - (1-target)*ln(1-output)``
 """
-struct CrossentropyLoss <: SupervisedLoss end
-const LogitProbLoss = CrossentropyLoss
+struct CrossEntropyLoss <: SupervisedLoss end
+const LogitProbLoss = CrossEntropyLoss
 
-function value(loss::CrossentropyLoss, target::Number, output::Number)
+function value(loss::CrossEntropyLoss, target::Number, output::Number)
     target >= 0 && target <=1 || error("target must be in [0,1]")
     output >= 0 && output <=1 || error("output must be in [0,1]")
     if target == 0
@@ -47,15 +47,15 @@ function value(loss::CrossentropyLoss, target::Number, output::Number)
         -(target * log(output) + (1-target) * log(1-output))
     end
 end
-deriv(loss::CrossentropyLoss, target::Number, output::Number) = (1-target) / (1-output) - target / output
-deriv2(loss::CrossentropyLoss, target::Number, output::Number) = (1-target) / (1-output)^2 + target / output^2
-value_deriv(loss::CrossentropyLoss, target::Number, output::Number) = (value(loss,target,output), deriv(loss,target,output))
-
-isdifferentiable(::CrossentropyLoss) = true
-isdifferentiable(::CrossentropyLoss, y, t) = t != 0 && t != 1
-istwicedifferentiable(::CrossentropyLoss) = true
-istwicedifferentiable(::CrossentropyLoss, y, t) = t != 0 && t != 1
-isconvex(::CrossentropyLoss) = true
+deriv(loss::CrossEntropyLoss, target::Number, output::Number) = (1-target) / (1-output) - target / output
+deriv2(loss::CrossEntropyLoss, target::Number, output::Number) = (1-target) / (1-output)^2 + target / output^2
+value_deriv(loss::CrossEntropyLoss, target::Number, output::Number) = (value(loss,target,output), deriv(loss,target,output))
+
+isdifferentiable(::CrossEntropyLoss) = true
+isdifferentiable(::CrossEntropyLoss, y, t) = t != 0 && t != 1
+istwicedifferentiable(::CrossEntropyLoss) = true
+istwicedifferentiable(::CrossEntropyLoss, y, t) = t != 0 && t != 1
+isconvex(::CrossEntropyLoss) = true
 
 # ===============================================================
 # L(target, output) = sign(agreement) < 0 ? 1 : 0

test/tst_loss.jl

@@ -484,7 +484,7 @@ end
 
 @testset "Test other loss against reference function" begin
     _crossentropyloss(y, t) = -y*log(t) - (1-y)*log(1-t)
-    test_value(CrossentropyLoss(), _crossentropyloss, 0:0.01:1, 0.01:0.01:0.99)
+    test_value(CrossEntropyLoss(), _crossentropyloss, 0:0.01:1, 0.01:0.01:0.99)
 
     _poissonloss(y, t) = exp(t) - t*y
     test_value(PoissonLoss(), _poissonloss, 0:10, range(0,stop=10,length=11))
@@ -540,8 +540,8 @@ end
 @testset "Test first and second derivatives of other losses" begin
     test_deriv(PoissonLoss(), -10:.2:10, 0:30)
     test_deriv2(PoissonLoss(), -10:.2:10, 0:30)
-    test_deriv(CrossentropyLoss(), 0:0.01:1, 0.01:0.01:0.99)
-    test_deriv2(CrossentropyLoss(), 0:0.01:1, 0.01:0.01:0.99)
+    test_deriv(CrossEntropyLoss(), 0:0.01:1, 0.01:0.01:0.99)
+    test_deriv2(CrossEntropyLoss(), 0:0.01:1, 0.01:0.01:0.99)
 end
 
 @testset "Test second derivatives of distance-based losses" begin