Use existing implementations in StatsFuns

Manifest.toml

@@ -3,19 +3,22 @@
 [[Base64]]
 uuid = "2a0f44e3-6c83-55bd-87e4-b1978d98bd5f"
 
+[[CompilerSupportLibraries_jll]]
+deps = ["Libdl", "Pkg"]
+git-tree-sha1 = "7c4f882c41faa72118841185afc58a2eb00ef612"
+uuid = "e66e0078-7015-5450-92f7-15fbd957f2ae"
+version = "0.3.3+0"
+
 [[Dates]]
 deps = ["Printf"]
 uuid = "ade2ca70-3891-5945-98fb-dc099432e06a"
 
-[[Distributed]]
-deps = ["Random", "Serialization", "Sockets"]
-uuid = "8ba89e20-285c-5b6f-9357-94700520ee1b"
-
 [[InteractiveUtils]]
 deps = ["Markdown"]
 uuid = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
 
 [[LibGit2]]
+deps = ["Printf"]
 uuid = "76f85450-5226-5b5a-8eaa-529ad045b433"
 
 [[Libdl]]
@@ -38,8 +41,14 @@ git-tree-sha1 = "c3d1a616362645754b18e12dbba96ec311b0867f"
 uuid = "a6bfbf70-4841-5cb9-aa18-3a8ad3c413ee"
 version = "2018.6.22+0"
 
+[[OpenSpecFun_jll]]
+deps = ["CompilerSupportLibraries_jll", "Libdl", "Pkg"]
+git-tree-sha1 = "d51c416559217d974a1113522d5919235ae67a87"
+uuid = "efe28fd5-8261-553b-a9e1-b2916fc3738e"
+version = "0.5.3+3"
+
 [[Pkg]]
-deps = ["Dates", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "REPL", "Random", "SHA", "Test", "UUIDs"]
+deps = ["Dates", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "REPL", "Random", "SHA", "UUIDs"]
 uuid = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 
 [[Printf]]
@@ -60,6 +69,18 @@ git-tree-sha1 = "d37400976e98018ee840e0ca4f9d20baa231dc6b"
 uuid = "ae029012-a4dd-5104-9daa-d747884805df"
 version = "1.0.1"
 
+[[Rmath]]
+deps = ["Random", "Rmath_jll"]
+git-tree-sha1 = "86c5647b565873641538d8f812c04e4c9dbeb370"
+uuid = "79098fc4-a85e-5d69-aa6a-4863f24498fa"
+version = "0.6.1"
+
+[[Rmath_jll]]
+deps = ["Libdl", "Pkg"]
+git-tree-sha1 = "1660f8fefbf5ab9c67560513131d4e933012fc4b"
+uuid = "f50d1b31-88e8-58de-be2c-1cc44531875f"
+version = "0.2.2+0"
+
 [[SHA]]
 uuid = "ea8e919c-243c-51af-8825-aaa63cd721ce"
 
@@ -73,13 +94,21 @@ uuid = "6462fe0b-24de-5631-8697-dd941f90decc"
 deps = ["LinearAlgebra", "Random"]
 uuid = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
 
+[[SpecialFunctions]]
+deps = ["OpenSpecFun_jll"]
+git-tree-sha1 = "d8d8b8a9f4119829410ecd706da4cc8594a1e020"
+uuid = "276daf66-3868-5448-9aa4-cd146d93841b"
+version = "0.10.3"
+
 [[Statistics]]
 deps = ["LinearAlgebra", "SparseArrays"]
 uuid = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 
-[[Test]]
-deps = ["Distributed", "InteractiveUtils", "Logging", "Random"]
-uuid = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
+[[StatsFuns]]
+deps = ["Rmath", "SpecialFunctions"]
+git-tree-sha1 = "04a5a8e6ab87966b43f247920eab053fd5fdc925"
+uuid = "4c63d2b9-4356-54db-8cca-17b64c39e42c"
+version = "0.9.5"
 
 [[UUIDs]]
 deps = ["Random", "SHA"]

Project.toml

@@ -9,9 +9,11 @@ NNPACK_jll = "a6bfbf70-4841-5cb9-aa18-3a8ad3c413ee"
 Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 Requires = "ae029012-a4dd-5104-9daa-d747884805df"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
+StatsFuns = "4c63d2b9-4356-54db-8cca-17b64c39e42c"
 
 [compat]
 Requires = "0.5, 1.0"
+StatsFuns = "0.9"
 julia = "1.3"
 
 [extras]

src/NNlib.jl

@@ -2,6 +2,7 @@ module NNlib
 using Pkg
 using Requires
 using NNPACK_jll
+using StatsFuns: logistic, softplus
 
 # Include APIs
 include("dim_helpers.jl")

src/activation.jl

@@ -12,10 +12,7 @@ export σ, sigmoid, hardσ, hardsigmoid, hardtanh, relu, leakyrelu, relu6, rrelu
 Classic [sigmoid](https://en.wikipedia.org/wiki/Sigmoid_function) activation
 function.
 """
-function σ(x::Real)
-    t = exp(-abs(x))
-    ifelse(x ≥ 0, inv(one(t) + t), t / (one(t) + t))
-end
+const σ = logistic
 const sigmoid = σ
 
 """
@@ -181,15 +178,6 @@ See [Quadratic Polynomials Learn Better Image Features](http://www.iro.umontreal
 """
 softsign(x::Real) = x / (one(x) + abs(x))
 
-
-"""
-    softplus(x) = log(exp(x) + 1)
-
-See [Deep Sparse Rectifier Neural Networks](http://proceedings.mlr.press/v15/glorot11a/glorot11a.pdf).
-"""
-softplus(x::Real) = ifelse(x > 0, x + log1p(exp(-x)), log1p(exp(x)))
-
-
 """
     logcosh(x)
 
@@ -222,7 +210,7 @@ See [Softshrink Activation Function](https://www.gabormelli.com/RKB/Softshrink_A
 softshrink(x::Real, λ = oftype(x/1, 0.5)) = min(max(zero(x), x - λ), x + λ)
 
 # Provide an informative error message if activation functions are called with an array
-for f in (:σ, :hardσ, :logσ, :hardtanh, :relu, :leakyrelu, :relu6, :rrelu, :elu, :gelu, :swish, :lisht, :selu, :celu, :trelu, :softsign, :softplus, :logcosh, :mish, :tanhshrink, :softshrink)
+for f in (:hardσ, :logσ, :hardtanh, :relu, :leakyrelu, :relu6, :rrelu, :elu, :gelu, :swish, :lisht, :selu, :celu, :trelu, :softsign, :logcosh, :mish, :tanhshrink, :softshrink)
     @eval $(f)(x::AbstractArray, args...) =
       error("Use broadcasting (`", $(string(f)), ".(x)`) to apply activation functions to arrays.")
 end

test/activation.jl

@@ -108,7 +108,9 @@ end
     @testset "Array input" begin
         x = rand(5)
         for a in ACTIVATION_FUNCTIONS
-            @test_throws ErrorException a(x)
+            if a != σ && a != softplus
+                @test_throws ErrorException a(x)
+            end
         end
     end