remove some v0.13 deprecations (#2493)

* remove v0.13 deprecations

Author: CarloLucibello

.buildkite/pipeline.yml

@@ -46,7 +46,7 @@ steps:
           using Pkg
           Pkg.resolve()'
     commands: |
-      printf "[MLDataDevices]\ngpu_backend = \"Metal\"\n" > LocalPreferences.toml
+      printf "[Flux]\ngpu_backend = \"Metal\"\n" > LocalPreferences.toml
 
     if: build.message !~ /\[skip tests\]/
     timeout_in_minutes: 60
@@ -63,7 +63,7 @@ steps:
   - label: "AMD GPU with Julia 1"
     plugins:
       - JuliaCI/julia#v1:
-          version: "1.10"
+          version: "1"
       - JuliaCI/julia-test#v1:
       - JuliaCI/julia-coverage#v1:
           dirs:
@@ -74,7 +74,7 @@ steps:
       rocm: "*"
       rocmgpu: "*"
     commands: |
-      printf "[MLDataDevices]\ngpu_backend = \"AMDGPU\"\n" > LocalPreferences.toml
+      printf "[Flux]\ngpu_backend = \"AMDGPU\"\n" > LocalPreferences.toml
     timeout_in_minutes: 60
     env:
       JULIA_AMDGPU_CORE_MUST_LOAD: "1"

.github/workflows/ci.yml

@@ -9,16 +9,23 @@ on:
       - master
     tags: '*'
 
+# needed for julia-actions/cache to delete old caches
+permissions:
+  actions: write
+  contents: read
+
 jobs:
   test:
-    name: Julia ${{ matrix.version }} - ${{ matrix.os }} - ${{ matrix.arch }} - ${{ github.event_name }}
+    name: Julia ${{ matrix.version }} - ${{ matrix.os }} - ${{ matrix.arch }}
     runs-on: ${{ matrix.os }}
     strategy:
       fail-fast: false
       matrix:
         version:
-          # - '1.9' # Uncomment when 1.10 is out. Replace this with the minimum Julia version that your package supports. 
-          - '1'
+          - '1.10' # Replace this with the minimum Julia version that your package supports. 
+          - '1'    # latest stable 1.x releas
+          # - 'pre'  # latest stable prerelease
+          - 'nightly' # latest nightly release
         os: [ubuntu-latest]
         arch: [x64]
         include:
@@ -29,7 +36,7 @@ jobs:
             version: '1'
             arch: aarch64
     steps:
-      - uses: actions/checkout@v4.2.1
+      - uses: actions/checkout@v4
       - uses: julia-actions/setup-julia@v2
         with:
           version: ${{ matrix.version }}
@@ -47,17 +54,16 @@ jobs:
       - uses: julia-actions/julia-buildpkg@v1
       - name: "Run test without coverage report"
         uses: julia-actions/julia-runtest@v1
-        if: ${{ !contains(fromJson('["1", "1.9"]'), matrix.version) || matrix.os != 'ubuntu-latest' }}
+        if: matrix.version != '1' || matrix.os != 'ubuntu-latest'
         with:
           coverage: false
-
       - name: "Run test with coverage report"
         uses: julia-actions/julia-runtest@v1
-        if: contains(fromJson('["1", "1.9"]'), matrix.version) && matrix.os == 'ubuntu-latest'
+        if: matrix.version == '1' && matrix.os == 'ubuntu-latest'
       - uses: julia-actions/julia-processcoverage@v1
-        if: contains(fromJson('["1", "1.9"]'), matrix.version) && matrix.os == 'ubuntu-latest'
+        if: matrix.version == '1' && matrix.os == 'ubuntu-latest'
       - uses: codecov/codecov-action@v4
-        if: contains(fromJson('["1", "1.9"]'), matrix.version) && matrix.os == 'ubuntu-latest'
+        if: matrix.version == '1' && matrix.os == 'ubuntu-latest'
         with:
           file: lcov.info
 

.github/workflows/nightly-ci.yml

@@ -67,4 +67,4 @@ SpecialFunctions = "2.1.2"
 Statistics = "1"
 Zygote = "0.6.67"
 cuDNN = "1"
-julia = "1.9"
+julia = "1.10"

Project.toml

@@ -6,7 +6,7 @@ import Flux
 import Flux: FluxCPUAdaptor, FluxAMDGPUAdaptor, _amd, adapt_storage, fmap
 import Flux: DenseConvDims, Conv, ConvTranspose, conv, conv_reshape_bias
 import NNlib
-
+using MLDataDevices: MLDataDevices
 using AMDGPU
 using Adapt
 using Random

ext/FluxAMDGPUExt/FluxAMDGPUExt.jl

@@ -10,6 +10,7 @@ using ChainRulesCore
 using Random
 using Adapt
 import Adapt: adapt_storage
+using MLDataDevices: MLDataDevices
 
 
 const USE_CUDA = Ref{Union{Nothing, Bool}}(nothing)

ext/FluxCUDAExt/FluxCUDAExt.jl

@@ -57,5 +57,5 @@ function _cuda(id::Union{Nothing, Int}, x)
 end
 
 function Flux._get_device(::Val{:CUDA}, id::Int)
-    return MLDataUtils.gpu_device(id+1, force=true)
+    return MLDataDevices.gpu_device(id+1, force=true)
 end

ext/FluxCUDAExt/functor.jl

@@ -16,10 +16,6 @@ _applyloss(loss, model, d...) = loss(model, d...)
 
 EnzymeRules.inactive(::typeof(Flux.Losses._check_sizes), args...) = true
 
-using Flux: _old_to_new  # from src/deprecations.jl
-train!(loss, model::Duplicated, data, opt::Optimise.AbstractOptimiser; cb=nothing) =
-  train!(loss, model, data, _old_to_new(opt); cb)
-
 function train!(loss, model::Duplicated, data, rule::Optimisers.AbstractRule; cb = nothing)
   train!(loss, model, data, _rule_to_state(model, rule); cb)
 end

ext/FluxEnzymeExt/FluxEnzymeExt.jl

@@ -4,7 +4,7 @@ import Flux
 import Flux: FluxCPUAdaptor, FluxMetalAdaptor, _metal, _isleaf, adapt_storage, fmap
 import NNlib
 using ChainRulesCore
-
+using MLDataDevices: MLDataDevices
 using Metal
 using Adapt
 using Random

ext/FluxMetalExt/FluxMetalExt.jl

@@ -35,6 +35,6 @@ end
 
 function Flux._get_device(::Val{:Metal}, id::Int)
     @assert id == 0 "Metal backend only supports one device at the moment"
-    return MLDataDevices.gpu_device()
+    return MLDataDevices.gpu_device(force=true)
 end