Skip to content

Instantly share code, notes, and snippets.

@hpoit

hpoit/regu_loss.jl

Created June 4, 2018 00:14
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save hpoit/938da6a5b048fe133aa9da2ad216e3fd to your computer and use it in GitHub Desktop.
Save hpoit/938da6a5b048fe133aa9da2ad216e3fd to your computer and use it in GitHub Desktop.
Download ZIP
regularizing loss
Raw
regu_loss.jl
julia> using Flux, Flux.Data.MNIST
julia> using Flux: onehotbatch, argmax, crossentropy, throttle
julia> using Base.Iterators: repeated
julia> # using CuArrays
# Classify MNIST digits with a simple multi-layer-perceptron
imgs = MNIST.images()
60000-element Array{Array{ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}},2},1}:
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
⋮
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}[Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); … ; Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0); Gray{N0f8}(0.0) Gray{N0f8}(0.0) … Gray{N0f8}(0.0) Gray{N0f8}(0.0)]
julia> # Stack images into one large batch
X = hcat(float.(reshape.(imgs, :))...) |> gpu
784×60000 Array{Float64,2}:
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
⋮ ⋮ ⋮ ⋱ ⋮ ⋮
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
julia> labels = MNIST.labels()
60000-element Array{Int64,1}:
5
0
4
1
9
2
1
3
1
4
3
⋮
8
9
2
9
5
1
8
3
5
6
8
julia> # One-hot-encode the labels
Y = onehotbatch(labels, 0:9) |> gpu
10×60000 Flux.OneHotMatrix{Array{Flux.OneHotVector,1}}:
false true false false false false false false false … false false false false false false false false
false false false true false false true false true false false true false false false false false
false false false false false true false false false false false false false false false false false
false false false false false false false true false false false false false true false false false
false false true false false false false false false false false false false false false false false
true false false false false false false false false … false true false false false true false false
false false false false false false false false false false false false false false false true false
false false false false false false false false false false false false false false false false false
false false false false false false false false false false false false true false false false true
false false false false true false false false false true false false false false false false false
julia> m = Chain(
Dense(28^2, 32, relu),
Dense(32, 10),
softmax) |> gpu
Chain(Dense(784, 32, NNlib.relu), Dense(32, 10), NNlib.softmax)
julia> penalty() = vecnorm(m.W) + vecnorm(m.b)
penalty (generic function with 1 method)
julia> loss(x, y) = crossentropy(m(x), y) + penalty()
loss (generic function with 1 method)
julia> loss(x, y) = crossentropy(m(x), y) + sum(vecnorm, params(m))
loss (generic function with 1 method)
julia> accuracy(x, y) = mean(argmax(m(x)) .== argmax(y))
accuracy (generic function with 1 method)
julia> dataset = repeated((X, Y), 200)
Base.Iterators.Take{Base.Iterators.Repeated{Tuple{Array{Float64,2},Flux.OneHotMatrix{Array{Flux.OneHotVector,1}}}}}(Base.Iterators.Repeated{Tuple{Array{Float64,2},Flux.OneHotMatrix{Array{Flux.OneHotVector,1}}}}(([0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0; … ; 0.0 0.0 … 0.0 0.0; 0.0 0.0 … 0.0 0.0], Bool[false true … false false; false false … false false; … ; false false … false true; false false … false false])), 200)
julia> evalcb = () -> @show(loss(X, Y))
(::#3) (generic function with 1 method)
julia> opt = ADAM(params(m))
(::#71) (generic function with 1 method)
julia> Flux.train!(loss, dataset, opt, cb = throttle(evalcb, 10))
loss(X, Y) = 13.83939362948925 (tracked)
loss(X, Y) = 11.445319293405092 (tracked)
loss(X, Y) = 9.428668991359263 (tracked)
loss(X, Y) = 7.748910424289779 (tracked)
loss(X, Y) = 6.405258815488118 (tracked)
loss(X, Y) = 5.495063048609246 (tracked)
loss(X, Y) = 5.071706004483157 (tracked)
loss(X, Y) = 4.768047350782004 (tracked)
loss(X, Y) = 4.493566747114699 (tracked)
loss(X, Y) = 4.265064443780441 (tracked)
loss(X, Y) = 4.05666850108693 (tracked)
loss(X, Y) = 3.8723977094343525 (tracked)
loss(X, Y) = 3.6842825789424154 (tracked)
loss(X, Y) = 3.5160295752796697 (tracked)
julia> accuracy(X, Y)
0.11308333333333333
julia> tX = hcat(float.(reshape.(MNIST.images(:test), :))...) |> gpu
784×10000 Array{Float64,2}:
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
⋮ ⋮ ⋮ ⋱ ⋮ ⋮
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 … 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
julia> tY = onehotbatch(MNIST.labels(:test), 0:9) |> gpu
10×10000 Flux.OneHotMatrix{Array{Flux.OneHotVector,1}}:
false false false true false false false false false … false true false false false false false false
false false true false false true false false false false false true false false false false false
false true false false false false false false false false false false true false false false false
false false false false false false false false false false false false false true false false false
false false false false true false true false false false false false false false true false false
false false false false false false false false true … false false false false false false true false
false false false false false false false false false false false false false false false false true
true false false false false false false false false false false false false false false false false
false false false false false false false false false false false false false false false false false
false false false false false false false true false true false false false false false false false
julia> accuracy(tX, tY)
0.1116
julia>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment