Change formatting style

.JuliaFormatter.toml

@@ -1,2 +1,2 @@
 style = "sciml"
-whitespace_in_kwargs = true
+whitespace_in_kwargs = false

.github/workflows/FormatCheck.yml

@@ -26,8 +26,11 @@ jobs:
         # This will use the latest version by default but you can set the version like so:
         #
         # julia  -e 'using Pkg; Pkg.add(PackageSpec(name="JuliaFormatter", version="0.13.0"))'
+        #
+        # FIXME: Before merging change to default release
+        # julia  -e 'using Pkg; Pkg.add(PackageSpec(name="JuliaFormatter"))'
         run: |
-          julia  -e 'using Pkg; Pkg.add(PackageSpec(name="JuliaFormatter"))'
+          julia  -e 'using Pkg; Pkg.add(PackageSpec(url="https://github.com/YingboMa/JuliaFormatter.jl.git", rev="myb/scimlstyle"))'
           julia  -e 'using JuliaFormatter; format(".", verbose=true)'
       - name: Format check
         run: |

.github/workflows/FormatPR.yml

@@ -9,7 +9,7 @@ jobs:
       - uses: actions/checkout@v2
       - name: Install JuliaFormatter and format
         run: |
-          julia  -e 'import Pkg; Pkg.add("JuliaFormatter")'
+          julia  -e 'using Pkg; Pkg.add(PackageSpec(url="https://github.com/YingboMa/JuliaFormatter.jl.git", rev="myb/scimlstyle"))'
           julia  -e 'using JuliaFormatter; format(".")'
       # https://github.com/marketplace/actions/create-pull-request
       # https://github.com/peter-evans/create-pull-request#reference-example

docs/make.jl

@@ -2,7 +2,7 @@ using Documenter, Lux, Literate, Pkg
 
 # Precompile example dependencies
 Pkg.activate(joinpath(@__DIR__, "..", "examples"))
-Pkg.develop(PackageSpec(; path = joinpath(@__DIR__, "..")))
+Pkg.develop(PackageSpec(; path=joinpath(@__DIR__, "..")))
 Pkg.instantiate()
 Pkg.precompile()
 Pkg.activate(@__DIR__)
@@ -13,8 +13,8 @@ if haskey(ENV, "GITHUB_ACTIONS")
 end
 
 deployconfig = Documenter.auto_detect_deploy_system()
-Documenter.post_status(deployconfig; type = "pending",
-                       repo = "github.com/avik-pal/Lux.jl.git")
+Documenter.post_status(deployconfig; type="pending",
+                       repo="github.com/avik-pal/Lux.jl.git")
 
 # Tutorials
 get_example_path(p) = joinpath(@__DIR__, "..", "examples", p)
@@ -28,13 +28,13 @@ ADVANCED_TUTORIALS = []
 ADVANCED_TUTORIAL_NAMES = []
 MAPPING = Dict("beginner" => [], "intermediate" => [], "advanced" => [])
 
-for (d, names, paths) in
-    (("beginner", BEGINNER_TUTORIAL_NAMES, BEGINNER_TUTORIALS),
-     ("intermediate", INTERMEDIATE_TUTORIAL_NAMES, INTERMEDIATE_TUTORIALS),
-     ("advanced", ADVANCED_TUTORIAL_NAMES, ADVANCED_TUTORIALS))
+for (d, names, paths) in (("beginner", BEGINNER_TUTORIAL_NAMES, BEGINNER_TUTORIALS),
+                          ("intermediate", INTERMEDIATE_TUTORIAL_NAMES,
+                           INTERMEDIATE_TUTORIALS),
+                          ("advanced", ADVANCED_TUTORIAL_NAMES, ADVANCED_TUTORIALS))
     for (n, p) in zip(names, paths)
         Literate.markdown(get_example_path(p), joinpath(OUTPUT, d, dirname(p));
-                          documenter = true)
+                          documenter=true)
         push!(MAPPING[d],
               n => joinpath("examples/generated", d, dirname(p),
                             splitext(basename(p))[1] * ".md"))
@@ -44,17 +44,17 @@ end
 display(MAPPING)
 
 makedocs(;
-         sitename = "Lux",
-         authors = "Avik Pal et al.",
-         clean = true,
-         doctest = false,
-         modules = [Lux],
-         format = Documenter.HTML(;
-                                  prettyurls = get(ENV, "CI", nothing) == "true",
-                                  assets = ["assets/custom.css"]
-                                  # analytics = "G-Q8GYTEVTZ2"
-                                  ),
-         pages = [
+         sitename="Lux",
+         authors="Avik Pal et al.",
+         clean=true,
+         doctest=false,
+         modules=[Lux],
+         format=Documenter.HTML(;
+                                prettyurls=get(ENV, "CI", nothing) == "true",
+                                assets=["assets/custom.css"]
+                                # analytics = "G-Q8GYTEVTZ2"
+                                ),
+         pages=[
              "Lux: Explicitly Parameterized Neural Networks" => "index.md",
              "Introduction" => [
                  "All about Lux" => "introduction/overview.md",
@@ -78,7 +78,7 @@ makedocs(;
              ],
          ])
 
-deploydocs(; repo = "github.com/avik-pal/Lux.jl.git", push_preview = true,
-           devbranch = "main")
+deploydocs(; repo="github.com/avik-pal/Lux.jl.git", push_preview=true,
+           devbranch="main")
 
 Pkg.activate(@__DIR__)

examples/BayesianNN/main.jl

@@ -52,8 +52,8 @@ function plot_data()
     x2 = first.(xt0s)
     y2 = last.(xt0s)
 
-    plt = Plots.scatter(x1, y1; color = "red", clim = (0, 1))
-    Plots.scatter!(plt, x2, y2; color = "blue", clim = (0, 1))
+    plt = Plots.scatter(x1, y1; color="red", clim=(0, 1))
+    Plots.scatter!(plt, x2, y2; color="blue", clim=(0, 1))
 
     return plt
 end
@@ -135,8 +135,8 @@ _, i = findmax(ch[:lp])
 i = i.I[1]
 
 ## Plot the posterior distribution with a contour plot
-x1_range = collect(range(-6; stop = 6, length = 25))
-x2_range = collect(range(-6; stop = 6, length = 25))
+x1_range = collect(range(-6; stop=6, length=25))
+x2_range = collect(range(-6; stop=6, length=25))
 Z = [nn_forward([x1, x2], theta[i, :])[1] for x1 in x1_range, x2 in x2_range]
 contour!(x1_range, x2_range, Z)
 
@@ -157,8 +157,8 @@ end
 plot_data()
 
 n_end = 1500
-x1_range = collect(range(-6; stop = 6, length = 25))
-x2_range = collect(range(-6; stop = 6, length = 25))
+x1_range = collect(range(-6; stop=6, length=25))
+x2_range = collect(range(-6; stop=6, length=25))
 Z = [nn_predict([x1, x2], theta, n_end)[1] for x1 in x1_range, x2 in x2_range]
 contour!(x1_range, x2_range, Z)
 
@@ -170,5 +170,5 @@ n_end = 1000
 anim = @gif for i in 1:n_end
     plot_data()
     Z = [nn_forward([x1, x2], theta[i, :])[1] for x1 in x1_range, x2 in x2_range]
-    contour!(x1_range, x2_range, Z; title = "Iteration $i", clim = (0, 1))
+    contour!(x1_range, x2_range, Z; title="Iteration $i", clim=(0, 1))
 end every 5

examples/ImageNet/main.jl

@@ -26,17 +26,17 @@ import DataLoaders: LearnBase                           # Extending Datasets
 import MLUtils
 
 # Distributed Training
-FluxMPI.Init(; verbose = true)
+FluxMPI.Init(; verbose=true)
 CUDA.allowscalar(false)
 
 # unsafe_free OneHotArrays
 CUDA.unsafe_free!(x::OneHotArray) = CUDA.unsafe_free!(x.indices)
 
 # Image Classification Models
-VGG11_BN(args...; kwargs...) = VGG11(args...; batchnorm = true, kwargs...)
-VGG13_BN(args...; kwargs...) = VGG13(args...; batchnorm = true, kwargs...)
-VGG16_BN(args...; kwargs...) = VGG16(args...; batchnorm = true, kwargs...)
-VGG19_BN(args...; kwargs...) = VGG19(args...; batchnorm = true, kwargs...)
+VGG11_BN(args...; kwargs...) = VGG11(args...; batchnorm=true, kwargs...)
+VGG13_BN(args...; kwargs...) = VGG13(args...; batchnorm=true, kwargs...)
+VGG16_BN(args...; kwargs...) = VGG16(args...; batchnorm=true, kwargs...)
+VGG19_BN(args...; kwargs...) = VGG19(args...; batchnorm=true, kwargs...)
 MobileNetv3_small(args...; kwargs...) = MobileNetv3(:small, args...; kwargs...)
 MobileNetv3_large(args...; kwargs...) = MobileNetv3(:large, args...; kwargs...)
 ResNeXt50(args...; kwargs...) = ResNeXt(50, args...; kwargs...)
@@ -75,7 +75,7 @@ AVAILABLE_IMAGENET_MODELS = [
 
 IMAGENET_MODELS_DICT = Dict(string(model) => model for model in AVAILABLE_IMAGENET_MODELS)
 
-function get_model(model_name::String, models_dict::Dict, rng, args...; warmup = true,
+function get_model(model_name::String, models_dict::Dict, rng, args...; warmup=true,
                    kwargs...)
     model = Lux.transform(models_dict[model_name](args...; kwargs...).layers)
     ps, st = Lux.setup(rng, model) .|> gpu
@@ -93,8 +93,8 @@ function get_model(model_name::String, models_dict::Dict, rng, args...; warmup =
     end
 
     if is_distributed()
-        ps = FluxMPI.synchronize!(ps; root_rank = 0)
-        st = FluxMPI.synchronize!(st; root_rank = 0)
+        ps = FluxMPI.synchronize!(ps; root_rank=0)
+        st = FluxMPI.synchronize!(st; root_rank=0)
         should_log() && println("$(now()) ==> models synced across all ranks")
     end
 
@@ -160,7 +160,7 @@ function parse_commandline_arguments()
 end
 
 # Loss Function
-logitcrossentropyloss(ŷ, y) = mean(-sum(y .* logsoftmax(ŷ; dims = 1); dims = 1))
+logitcrossentropyloss(ŷ, y) = mean(-sum(y .* logsoftmax(ŷ; dims=1); dims=1))
 
 function logitcrossentropyloss(x, y, model, ps, st)
     ŷ, st_ = model(x, ps, st)
@@ -181,10 +181,10 @@ end
 update_lr(st_opt::NamedTuple, eta) = fmap(l -> update_lr(l, eta), st_opt)
 
 # Accuracy
-function accuracy(ŷ, y, topk = (1,))
+function accuracy(ŷ, y, topk=(1,))
     maxk = maximum(topk)
 
-    pred_labels = partialsortperm.(eachcol(ŷ), (1:maxk,), rev = true)
+    pred_labels = partialsortperm.(eachcol(ŷ), (1:maxk,), rev=true)
     true_labels = onecold(y)
 
     accuracies = Vector{Float32}(undef, length(topk))
@@ -202,11 +202,11 @@ is_distributed() = FluxMPI.Initialized() && total_workers() > 1
 should_log() = !FluxMPI.Initialized() || local_rank() == 0
 
 # Checkpointing
-function save_checkpoint(state, is_best, filename = "checkpoint.pth.tar")
+function save_checkpoint(state, is_best, filename="checkpoint.pth.tar")
     if should_log()
         serialize(filename, state)
         if is_best
-            cp(filename, "model_best.pth.tar"; force = true)
+            cp(filename, "model_best.pth.tar"; force=true)
         end
     end
 end
@@ -260,12 +260,12 @@ function ImageDataset(folder::String, augmentation_pipeline, normalization_param
             "n02105855_2933.JPEG",
         ]
         remove_files = joinpath.((folder,),
-                                 joinpath.(first.(rsplit.(remove_files, "_", limit = 2)),
+                                 joinpath.(first.(rsplit.(remove_files, "_", limit=2)),
                                            remove_files))
 
         image_files = [setdiff(Set(image_files), Set(remove_files))...]
 
-        labels = [mapping[x] for x in map(x -> x[2], rsplit.(image_files, "/", limit = 3))]
+        labels = [mapping[x] for x in map(x -> x[2], rsplit.(image_files, "/", limit=3))]
     else
         vallist = hcat(split.(readlines(joinpath(@__DIR__, "val_list.txt")))...)
         labels = parse.(Int, vallist[2, :]) .+ 1
@@ -315,7 +315,7 @@ end
 
 function AverageMeter(name::String, fmt::String)
     fmtstr = FormatExpr("$name {1:$fmt} ({2:$fmt})")
-    return AverageMeter(; fmtstr = fmtstr)
+    return AverageMeter(; fmtstr=fmtstr)
 end
 
 function update!(meter::AverageMeter, val, n::Int)
@@ -333,7 +333,7 @@ struct ProgressMeter{N}
     meters::NTuple{N, AverageMeter}
 end
 
-function ProgressMeter(num_batches::Int, meters::NTuple{N}, prefix::String = "") where {N}
+function ProgressMeter(num_batches::Int, meters::NTuple{N}, prefix::String="") where {N}
     fmt = "%" * string(length(string(num_batches))) * "d"
     prefix = prefix != "" ? endswith(prefix, " ") ? prefix : prefix * " " : ""
     batch_fmtstr = generate_formatter("$prefix[$fmt/" * sprintf1(fmt, num_batches) * "]")
@@ -464,11 +464,11 @@ function main(args)
             println("$(now()) => creating model `$(args["arch"])`")
         end
     end
-    model, ps, st = get_model(args["arch"], IMAGENET_MODELS_DICT, rng; warmup = true,
-                              pretrain = args["pretrained"])
+    model, ps, st = get_model(args["arch"], IMAGENET_MODELS_DICT, rng; warmup=true,
+                              pretrain=args["pretrained"])
 
-    normalization_parameters = (mean = reshape([0.485f0, 0.456f0, 0.406f0], 1, 1, 3),
-                                std = reshape([0.229f0, 0.224f0, 0.225f0], 1, 1, 3))
+    normalization_parameters = (mean=reshape([0.485f0, 0.456f0, 0.406f0], 1, 1, 3),
+                                std=reshape([0.229f0, 0.224f0, 0.225f0], 1, 1, 3))
     train_data_augmentation = Resize(256, 256) |> FlipX(0.5) |> RCropSize(224, 224)
     val_data_augmentation = Resize(256, 256) |> CropSize(224, 224)
     train_dataset = ImageDataset(joinpath(args["data"], "train"),
@@ -496,7 +496,7 @@ function main(args)
         optimiser_state = FluxMPI.synchronize!(optimiser_state)
         should_log() && println("$(now()) ==> synced optimiser state across all ranks")
     end
-    scheduler = Step(λ = args["learning-rate"], γ = 0.1f0, step_sizes = 30)
+    scheduler = Step(λ=args["learning-rate"], γ=0.1f0, step_sizes=30)
 
     if args["resume"] != ""
         if isfile(args["resume"])

examples/NeuralODE/main.jl

@@ -26,13 +26,13 @@ function loadmnist(batchsize, train_split)
     ## Process images into (H,W,C,BS) batches
     x_data = Float32.(reshape(imgs, size(imgs, 1), size(imgs, 2), 1, size(imgs, 3)))
     y_data = onehot(labels_raw)
-    (x_train, y_train), (x_test, y_test) = splitobs((x_data, y_data); at = train_split)
+    (x_train, y_train), (x_test, y_test) = splitobs((x_data, y_data); at=train_split)
 
     return (
             ## Use DataLoader to automatically minibatch and shuffle the data
-            DataLoader(collect.((x_train, y_train)); batchsize = batchsize, shuffle = true),
+            DataLoader(collect.((x_train, y_train)); batchsize=batchsize, shuffle=true),
             ## Don't shuffle the test data
-            DataLoader(collect.((x_test, y_test)); batchsize = batchsize, shuffle = false))
+            DataLoader(collect.((x_test, y_test)); batchsize=batchsize, shuffle=false))
 end
 
 # ## Define the Neural ODE Layer
@@ -49,9 +49,9 @@ struct NeuralODE{M <: Lux.AbstractExplicitLayer, So, Se, T, K} <:
 end
 
 function NeuralODE(model::Lux.AbstractExplicitLayer;
-                   solver = Tsit5(),
-                   sensealg = InterpolatingAdjoint(; autojacvec = ZygoteVJP()),
-                   tspan = (0.0f0, 1.0f0),
+                   solver=Tsit5(),
+                   sensealg=InterpolatingAdjoint(; autojacvec=ZygoteVJP()),
+                   tspan=(0.0f0, 1.0f0),
                    kwargs...)
     return NeuralODE(model, solver, sensealg, tspan, kwargs)
 end
@@ -62,13 +62,13 @@ function (n::NeuralODE)(x, ps, st)
         return u_
     end
     prob = ODEProblem{false}(ODEFunction{false}(dudt), x, n.tspan, ps)
-    return solve(prob, n.solver; sensealg = n.sensealg, n.kwargs...), st
+    return solve(prob, n.solver; sensealg=n.sensealg, n.kwargs...), st
 end
 
 function diffeqsol_to_array(x::ODESolution{T, N, <:AbstractVector{<:CuArray}}) where {T, N}
-    dropdims(gpu(x); dims = 3)
+    dropdims(gpu(x); dims=3)
 end
-diffeqsol_to_array(x::ODESolution) = dropdims(Array(x); dims = 3)
+diffeqsol_to_array(x::ODESolution) = dropdims(Array(x); dims=3)
 
 # ## Create and Initialize the Neural ODE Layer
 function create_model()
@@ -77,10 +77,10 @@ function create_model()
                   Dense(784, 20, tanh),
                   NeuralODE(Chain(Dense(20, 10, tanh), Dense(10, 10, tanh),
                                   Dense(10, 20, tanh));
-                            save_everystep = false,
-                            reltol = 1.0f-3,
-                            abstol = 1.0f-3,
-                            save_start = false),
+                            save_everystep=false,
+                            reltol=1.0f-3,
+                            abstol=1.0f-3,
+                            save_start=false),
                   diffeqsol_to_array,
                   Dense(20, 10))
 
@@ -97,7 +97,7 @@ end
 # ## Define Utility Functions
 get_class(x) = argmax.(eachcol(x))
 
-logitcrossentropy(ŷ, y) = mean(-sum(y .* logsoftmax(ŷ); dims = 1))
+logitcrossentropy(ŷ, y) = mean(-sum(y .* logsoftmax(ŷ); dims=1))
 
 function loss(x, y, model, ps, st)
     ŷ, st = model(x, ps, st)

examples/SimpleRNN/main.jl

@@ -16,7 +16,7 @@ using MLUtils, Optimisers, Zygote, NNlib, Random, Statistics
 
 # We will use MLUtils to generate 500 (noisy) clockwise and 500 (noisy) anticlockwise spirals. Using this data we will create a `MLUtils.DataLoader`. Our dataloader will give us sequences of size 2 × seq_len × batch_size and we need to predict a binary value whether the sequence is clockwise or anticlockwise
 
-function get_dataloaders(; dataset_size = 1000, sequence_length = 50)
+function get_dataloaders(; dataset_size=1000, sequence_length=50)
     ## Create the spirals
     data = [MLUtils.Datasets.make_spiral(sequence_length) for _ in 1:dataset_size]
     ## Get the labels
@@ -25,16 +25,16 @@ function get_dataloaders(; dataset_size = 1000, sequence_length = 50)
                          for d in data[1:(dataset_size ÷ 2)]]
     anticlockwise_spirals = [reshape(d[1][:, (sequence_length + 1):end], :, sequence_length,
                                      1) for d in data[((dataset_size ÷ 2) + 1):end]]
-    x_data = Float32.(cat(clockwise_spirals..., anticlockwise_spirals...; dims = 3))
+    x_data = Float32.(cat(clockwise_spirals..., anticlockwise_spirals...; dims=3))
     ## Split the dataset
-    (x_train, y_train), (x_val, y_val) = splitobs((x_data, labels); at = 0.8,
-                                                  shuffle = true)
+    (x_train, y_train), (x_val, y_val) = splitobs((x_data, labels); at=0.8,
+                                                  shuffle=true)
     ## Create DataLoaders
     return (
             ## Use DataLoader to automatically minibatch and shuffle the data
-            DataLoader(collect.((x_train, y_train)); batchsize = 128, shuffle = true),
+            DataLoader(collect.((x_train, y_train)); batchsize=128, shuffle=true),
             ## Don't shuffle the validation data
-            DataLoader(collect.((x_val, y_val)); batchsize = 128, shuffle = false))
+            DataLoader(collect.((x_val, y_val)); batchsize=128, shuffle=false))
 end
 
 # ## Creating a Classifier
@@ -72,7 +72,7 @@ function (s::SpiralClassifier)(x::AbstractArray{T, 3}, ps::NamedTuple,
     ## After running through the sequence we will pass the output through the classifier
     y, st_classifier = s.classifier(h, ps.classifier, st.classifier)
     ## Finally remember to create the updated state
-    st = merge(st, (classifier = st_classifier, lstm_cell = st_lstm))
+    st = merge(st, (classifier=st_classifier, lstm_cell=st_lstm))
     return vec(y), st
 end