Use new public feature (#2342)

Project.toml

@@ -5,6 +5,7 @@ version = "0.14.6"
 [deps]
 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
 ChainRulesCore = "d360d2e6-b24c-11e9-a2a3-2a2ae2dbcce4"
+Compat = "34da2185-b29b-5c13-b0c7-acf172513d20"
 Functors = "d9f16b24-f501-4c13-a1f2-28368ffc5196"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 MLUtils = "f1d291b0-491e-4a28-83b9-f70985020b54"
@@ -38,6 +39,7 @@ AMDGPU = "0.6, 0.7"
 Adapt = "3.0"
 CUDA = "4, 5"
 ChainRulesCore = "1.12"
+Compat = "4.10.0"
 Functors = "0.4"
 MLUtils = "0.4"
 MacroTools = "0.5"

src/Flux.jl

@@ -1,6 +1,7 @@
 module Flux
 
 using Base: tail
+using Compat: @compat # for @compat public
 using Preferences
 using LinearAlgebra, Statistics, Random  # standard lib
 using MacroTools, Reexport, ProgressLogging, SpecialFunctions
@@ -30,6 +31,15 @@ export Chain, Dense, Embedding, Maxout, SkipConnection, Parallel, PairwiseFusion
        fmap, cpu, gpu, f32, f64, f16, rand32, randn32, zeros32, ones32,
        testmode!, trainmode!
 
+@compat(public, ( # mark unexported symbols as API, on Julia 1.11
+  # modules
+  Losses,
+  # layers
+  Bilinear, Scale, dropout,
+  # utils
+  outputsize, state,
+))
+
 include("optimise/Optimise.jl")
 using .Optimise
 export Descent, Adam, Momentum, Nesterov, RMSProp,
@@ -47,6 +57,15 @@ using Adapt, Functors, OneHotArrays
 include("utils.jl")
 include("functor.jl")
 
+@compat(public, (
+  # from OneHotArrays.jl
+  onehot, onehotbatch, onecold,  
+  # from Functors.jl
+  functor, @functor,
+  # from Optimise/Train/Optimisers.jl
+  setup, update!, destructure, freeze!, adjust!, params, trainable
+))
+
 # Pirate error to catch a common mistake.
 Functors.functor(::Type{<:MLUtils.DataLoader}, x) = error("`DataLoader` does not support Functors.jl, thus functions like `Flux.gpu` will not act on its contents.")
 
@@ -69,5 +88,37 @@ include("deprecations.jl")
 include("losses/Losses.jl")
 using .Losses
 
+@compat(public, (
+  # init
+  glorot_uniform,
+  glorot_normal,
+  kaiming_uniform,
+  kaiming_normal,
+  truncated_normal,
+  orthogonal,
+  sparse_init,
+  identity_init,
+
+  # Losses
+  binary_focal_loss,
+  binarycrossentropy,
+  crossentropy,
+  dice_coeff_loss,
+  focal_loss,
+  hinge_loss,
+  huber_loss,
+  kldivergence,
+  label_smoothing,
+  logitbinarycrossentropy,
+  logitcrossentropy,
+  mae,
+  mse,
+  msle,
+  poisson_loss,
+  siamese_contrastive_loss,
+  squared_hinge_loss,
+  tversky_loss,
+))
+
 
 end # module