WeightNormWeight is now called WeightNormParam

WeightNorm for several params, single dim

Test for Scalar and Vector dims

Test newly created WN equality

Simplified some bits

Missing last constructor

Author: bhvieira

src/Flux.jl

@@ -11,7 +11,7 @@ export gradient
 
 export Chain, Dense, Maxout, RNN, LSTM, GRU, Conv, CrossCor, ConvTranspose, MaxPool, MeanPool,
        DepthwiseConv, Dropout, AlphaDropout, LayerNorm, BatchNorm, InstanceNorm, GroupNorm,
-       WeightNorm, WeightNormWeight, SkipConnection, params, fmap, cpu, gpu, f32, f64
+       WeightNorm, WeightNormParam, SkipConnection, params, fmap, cpu, gpu, f32, f64
 
 include("optimise/Optimise.jl")
 using .Optimise

src/layers/normalise.jl

@@ -371,13 +371,13 @@ end
 Weight Normalization.
 This layer reparametrizes weights (w) of a layer with its decomposition into magnitude (g) and direction (v).
 
-	WeightNorm(layer, weight::Union{Symbol,Int}, dim)
+	WeightNorm(layer, weight, dim)
 
 ``layer`` is the layer being normalized.
 
-``weight`` is the parameter to be normalized.
+``weight`` are the parameters to be normalized.
 
-``dim`` is the dimension of normalization.
+``dim`` are the dimension of normalization.
 Often, its the dimension encoding the output channels.
 
 Example:
@@ -390,55 +390,62 @@ wndB = WeightNorm(d, :W, 1:2); #Now we normalize all directions together, keepin
 Link : https://arxiv.org/pdf/1602.07868.pdf
 """
 
-struct WeightNormWeight{T,N,I}
+struct WeightNormParam{T,N,I}
   g::AbstractArray{T,N}
   v::AbstractArray{T,N}
   dim::I
 end
 
-Base.size(w::WeightNormWeight, i...) = size(w.v, i...)
-Base.size(w::WeightNormWeight) = size(w.v)
-Base.iterate(w::WeightNormWeight, i...) = iterate(w.g .* w.v ./ WN_mag(w.v, w.dim), i...)
-Base.getindex(w::WeightNormWeight, i...) = getindex(w.g .* w.v ./ WN_mag(w.v, w.dim), i...)
-Base.ndims(w::WeightNormWeight) = ndims(w.v)
+Base.size(w::WeightNormParam, i...) = size(w.v, i...)
+Base.size(w::WeightNormParam) = size(w.v)
+Base.iterate(w::WeightNormParam, i...) = iterate(w.g .* w.v ./ WN_mag(w.v, w.dim), i...)
+Base.getindex(w::WeightNormParam, i...) = getindex(w.g .* w.v ./ WN_mag(w.v, w.dim), i...)
+Base.ndims(w::WeightNormParam) = ndims(w.v)
+Base.length(w::WeightNormParam) = length(w.v)
 
-Flux.@functor WeightNormWeight
+@functor WeightNormParam
 
-WN_mag(p, dim) = sqrt.(sum(abs2.(p), dims = dim))
-WN_dir(p, mag, eps) = p ./ (mag .+ eps)
-WN_dir(p, mag) = WN_dir(p, mag, eps(eltype(p)))
+WN_mag(p, dim, eps) = sqrt.(sum(abs2.(p), dims = dim)) .+ eps
+WN_mag(p, dim) = WN_mag(p, dim, eps(eltype(p)))
+WN_dir(p, mag) = p ./ mag
 
 import Base.*, Base./, Base.+, Base.-
 for f in (:+, :-, :*, :/)
-  @eval ($f)(z::AbstractArray, w::WeightNormWeight) = ($f)(z, w.g .* w.v ./ WN_mag(w.v, w.dim))
-  @eval ($f)(w::WeightNormWeight, z::AbstractArray) = ($f)(w.g .* w.v ./ WN_mag(w.v, w.dim), z)
+  @eval ($f)(z::AbstractArray, w::WeightNormParam) = ($f)(z, w.g .* w.v ./ WN_mag(w.v, w.dim))
+  @eval ($f)(w::WeightNormParam, z::AbstractArray) = ($f)(w.g .* w.v ./ WN_mag(w.v, w.dim), z)
 end
 
-struct WeightNorm{L,E,I,W}
+struct WeightNorm{L}
   layer::L
-  eps::E
-  weight::W
-  dim::I
+  eps::Number
+  weight::Vector
+  dim::Vector
 end
 
-Flux.@functor WeightNorm
+@functor WeightNorm
 
 function Base.show(io::IO, wn::WeightNorm)
   print(io, "WeightNorm(", wn.layer, ", ", wn.weight, ", ", wn.dim, ")")
 end
 
-function WeightNorm(layer, weight::Union{Symbol,Int}, dim)
+function WeightNorm(layer, weight::Vector, dim::Vector)
   #Expose layer fields and constructor
   func, re = Flux.functor(layer)
   #Get the fields
   par = [getfield(layer, fn) for fn in keys(func)]
-  w = getfield(layer, weight)
-  g = WN_mag(w, dim)
-  v = WN_dir(w, g)
-  par[findfirst(keys(func) .== weight)] = WeightNormWeight(g, v, dim)
+  w = map(weight) do W
+      getfield(layer, W)
+  end
+  g = map((W, D) -> WN_mag(W, D), w, dim)
+  v = map((W, G) -> WN_dir(W, G), w, g)
+  par[indexin(weight,collect(keys(func)))] = WeightNormParam.(g, v, dim)
   return WeightNorm(re(par), eps(Float32), weight, dim)
 end
 
+WeightNorm(layer, weight::Symbol, dim::Vector) = WeightNorm(layer, [weight], dim)
+WeightNorm(layer, weight::Symbol, dim::Integer) = WeightNorm(layer, [weight], [dim])
+WeightNorm(layer, weight::Vector, dim::Integer) = WeightNorm(layer, weight, [dim for _ in axes(weight,1)])
+
 function (wn::WeightNorm)(x)
   wn.layer(x)
 end

test/layers/normalisation.jl

@@ -196,20 +196,26 @@ end
     d = Dense(10, 9, tanh)
     gs = gradient(() -> sum(abs2, d(fake_data)), params(d))
     W = d.W
-    for WN_dim in [1, 2, 1:2]
+    for WN_dim in [[1], 1, [2], 2, [1:2]]
       wnd = WeightNorm(d, :W, WN_dim)
       gswn = gradient(() -> sum(abs2, wnd(fake_data)), params(wnd))
       g = wnd.layer.W.g
       v = wnd.layer.W.v
-      normv = sum(abs2, v, dims = WN_dim)
 
       ΔW = gs[W]
       Δg = gswn[g]
       Δv = gswn[v]
-      @test sum(ΔW .* v ./ normv, dims = WN_dim) ≈ Δg
+      @test wnd(fake_data) ≈ d(fake_data)
+      if isa(WN_dim, Int)
+        normv = sum(abs2, v, dims = WN_dim)
+        @test sum(ΔW .* v ./ normv, dims = WN_dim) ≈ Δg
+      else
+        normv = sum(abs2, v, dims = WN_dim[1])
+        @test sum(ΔW .* v ./ normv, dims = WN_dim[1]) ≈ Δg
+      end
       @test g ./ normv .* ΔW - g .* Δg .* v ./ (normv.^2) ≈ Δv
       @test size(Δv) == size(ΔW)
-      @test isa(wnd.layer.W, WeightNormWeight)
+      @test isa(wnd.layer.W, Flux.WeightNormParam)
     end
   end
 end