ggml-python 0.0.37__tar.gz → 0.0.39__tar.gz

ggml_python-0.0.39/.github/workflows/lint.yaml

@@ -0,0 +1,29 @@
+name: Lint
+
+on:
+  pull_request:
+    branches:
+      - main
+  push:
+    branches:
+      - main
+
+jobs:
+  ruff:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install Ruff
+        run: python -m pip install "ruff>=0.15.7"
+
+      - name: Lint with Ruff
+        run: python -m ruff check ggml tests
+
+      - name: Check formatting with Ruff
+        run: python -m ruff format --check --exclude ggml/ggml.py ggml tests

ggml_python-0.0.39/.github/workflows/test.yaml

@@ -0,0 +1,158 @@
+name: Tests
+
+on:
+  pull_request:
+    branches:
+      - main
+  push:
+    branches:
+      - main
+
+jobs:
+  build-linux:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12"]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: "pip"
+      - name: Install dependencies
+        run: |
+          python3 -m pip install --upgrade pip pytest cmake scikit-build setuptools
+          python3 -m pip install \
+            --verbose \
+            --config-settings cmake.args='-DCMAKE_BUILD_TYPE=Debug;-DCMAKE_CXX_FLAGS=-g3;-DCMAKE_C_FLAGS=-g3' \
+            --config-settings cmake.verbose=true \
+            --config-settings logging.level=INFO \
+            --config-settings install.strip=false \
+            --editable .
+      - name: Test with pytest
+        run: |
+          python -m pytest -s -vvvv
+
+  build-windows:
+    runs-on: windows-latest
+    strategy:
+      matrix:
+        python-version: ["3.8", "3.9", "3.10", "3.11", "3.12"]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: "pip"
+
+      - name: Install dependencies
+        run: |
+          python3 -m pip install --upgrade pip pytest cmake scikit-build setuptools
+          python3 -m pip install --verbose --config-settings cmake.args='-DCMAKE_BUILD_TYPE=Debug;-DCMAKE_CXX_FLAGS=-g3;-DCMAKE_C_FLAGS=-g3' --config-settings cmake.verbose=true --config-settings logging.level=INFO --config-settings install.strip=false --editable .
+      - name: Test with pytest
+        run: |
+          python -m pytest -s -vvvv
+
+  build-macos:
+    runs-on: macos-15
+    strategy:
+      matrix:
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: "pip"
+      - name: System Info
+        run: |
+          uname -a
+          sysctl -n machdep.cpu.brand_string
+          python3 -c "import platform; print(platform.machine(), platform.architecture())"
+      - name: Install dependencies
+        run: |
+          python3 -m pip install --upgrade pip pytest cmake scikit-build setuptools
+          python3 -m pip install \
+            --verbose \
+            --config-settings cmake.args='-DCMAKE_BUILD_TYPE=Debug;-DCMAKE_CXX_FLAGS=-g3;-DCMAKE_C_FLAGS=-g3;-DGGML_NATIVE=off' \
+            --config-settings cmake.verbose=true \
+            --config-settings logging.level=INFO \
+            --config-settings install.strip=false \
+            --editable .
+      - name: Test with pytest
+        run: |
+          python -m pytest -s -vvvv
+
+  build-macos-intel:
+    runs-on: macos-15-intel
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+      - name: Set up Python 3.9
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.9"
+          cache: "pip"
+      - name: System Info
+        run: |
+          uname -a
+          sysctl -n machdep.cpu.brand_string
+          python3 -c "import platform; print(platform.machine(), platform.architecture())"
+      - name: Install dependencies
+        run: |
+          python3 -m pip install --upgrade pip pytest cmake scikit-build setuptools
+          python3 -m pip install \
+            --verbose \
+            --config-settings cmake.args='-DCMAKE_BUILD_TYPE=Debug;-DCMAKE_CXX_FLAGS=-g3;-DCMAKE_C_FLAGS=-g3;-DGGML_NATIVE=off' \
+            --config-settings cmake.verbose=true \
+            --config-settings logging.level=INFO \
+            --config-settings install.strip=false \
+            --editable .
+      - name: Test with pytest
+        run: |
+          python -m pytest -s -vvvv
+
+  build-macos-metal:
+    runs-on: macos-15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.9"
+          cache: "pip"
+      - name: System Info
+        run: |
+          uname -a
+          sysctl -n machdep.cpu.brand_string
+          python3 -c "import platform; print(platform.machine(), platform.architecture())"
+      - name: Install dependencies
+        run: |
+          python3 -m pip install --upgrade pip pytest cmake scikit-build setuptools
+          python3 -m pip install \
+            --verbose \
+            --config-settings cmake.args='-DCMAKE_BUILD_TYPE=Debug;-DCMAKE_CXX_FLAGS=-g3;-DCMAKE_C_FLAGS=-g3;-DGGML_NATIVE=off;-DGGML_METAL=On' \
+            --config-settings cmake.verbose=true \
+            --config-settings logging.level=INFO \
+            --config-settings install.strip=false \
+            --editable .
+      - name: Test with pytest
+        run: |
+          python -m pytest -s -vvvv

ggml_python-0.0.39/.github/workflows/wheels-cuda.yaml

@@ -0,0 +1,234 @@
+name: Wheels CUDA
+
+on: workflow_dispatch
+
+permissions:
+  contents: write
+
+jobs:
+  define_matrix:
+    name: Define Build Matrix
+    runs-on: ubuntu-22.04
+    outputs:
+      matrix: ${{ steps.set-matrix.outputs.matrix }}
+    defaults:
+      run:
+        shell: pwsh
+
+    steps:
+      - name: Define Job Output
+        id: set-matrix
+        run: |
+          $matrix = @{
+              'os' = @('ubuntu-22.04', 'windows-2022')
+              # wheel.py-api = "py3" makes the CUDA wheel interpreter-agnostic,
+              # so one builder per toolkit version is sufficient.
+              'pyver' = @("3.9")
+              'cuda' = @("11.8.0", "12.1.1", "12.2.2", "12.3.2", "12.4.1", "12.5.1", "13.0.2", "13.2.1")
+              'releasetag' = @("basic")
+              'exclude' = @(
+                @{ 'os' = 'windows-2022'; 'cuda' = '12.1.1' },
+                @{ 'os' = 'windows-2022'; 'cuda' = '12.2.2' },
+                @{ 'os' = 'windows-2022'; 'cuda' = '12.3.2' }
+              )
+          }
+
+          $matrixOut = ConvertTo-Json $matrix -Compress
+          Write-Output ('matrix=' + $matrixOut) >> $env:GITHUB_OUTPUT
+
+  build_wheels:
+    name: Build Wheel ${{ matrix.os }} ${{ matrix.pyver }} ${{ matrix.cuda }} ${{ matrix.releasetag == 'wheels' && 'AVX2' || matrix.releasetag }}
+    needs: define_matrix
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix: ${{ fromJSON(needs.define_matrix.outputs.matrix) }}
+    defaults:
+      run:
+        shell: pwsh
+    env:
+      CUDAVER: ${{ matrix.cuda }}
+      AVXVER: ${{ matrix.releasetag }}
+
+    steps:
+      - name: Set up MSVC for CUDA 11.8
+        if: runner.os == 'Windows' && matrix.cuda == '11.8.0'
+        uses: ilammy/msvc-dev-cmd@v1
+        with:
+          arch: x64
+          toolset: 14.29
+
+      - name: Set up MSVC
+        if: runner.os == 'Windows' && matrix.cuda != '11.8.0'
+        uses: ilammy/msvc-dev-cmd@v1
+        with:
+          arch: x64
+
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.pyver }}
+          cache: "pip"
+
+      - name: Setup Mamba
+        uses: conda-incubator/setup-miniconda@v3.1.0
+        with:
+          activate-environment: "ggml"
+          python-version: ${{ matrix.pyver }}
+          miniforge-version: latest
+          add-pip-as-python-dependency: true
+          auto-activate-base: false
+
+      - name: Install Dependencies
+        env:
+          MAMBA_DOWNLOAD_FAILFAST: "0"
+          MAMBA_NO_LOW_SPEED_LIMIT: "1"
+        run: |
+          $cudaVersion = $env:CUDAVER
+          $cudaChannel = "nvidia/label/cuda-$cudaVersion"
+          if ($cudaVersion -eq '11.8.0') {
+            if ($IsLinux) {
+              $cudaPackages = @(
+                "${cudaChannel}::cuda-nvcc_linux-64=11.8.0",
+                "${cudaChannel}::cuda-cccl=11.8.89",
+                "${cudaChannel}::cuda-cudart=11.8.89",
+                "${cudaChannel}::cuda-cudart-dev=11.8.89",
+                "${cudaChannel}::cuda-driver-dev=11.8.89",
+                "${cudaChannel}::libcublas=11.11.3.6",
+                "${cudaChannel}::libcublas-dev=11.11.3.6"
+              )
+            } elseif ($IsWindows) {
+              $cudaPackages = @(
+                "${cudaChannel}::cuda-nvcc_win-64=11.8.0",
+                "${cudaChannel}::cuda-cccl=11.8.89",
+                "${cudaChannel}::cuda-cudart=11.8.89",
+                "${cudaChannel}::cuda-cudart-dev=11.8.89",
+                "${cudaChannel}::libcublas=11.11.3.6",
+                "${cudaChannel}::libcublas-dev=11.11.3.6"
+              )
+            } else {
+              throw 'Unsupported CUDA wheel build platform'
+            }
+            mamba install -y --channel-priority flexible --override-channels -c $cudaChannel $cudaPackages
+          } elseif ($IsLinux) {
+            mamba install -y --channel-priority flexible --override-channels -c $cudaChannel "${cudaChannel}::cuda-toolkit=$cudaVersion" "${cudaChannel}::cuda-nvcc_linux-64" "${cudaChannel}::cuda-cccl" "${cudaChannel}::cuda-cudart" "${cudaChannel}::cuda-cudart-dev"
+          } elseif ($IsWindows) {
+            if ($cudaVersion -like '12.5.*' -or [version]$cudaVersion -ge [version]"13.0") {
+              # The Windows 12.5+ toolkit meta-package pulls compiler activation
+              # scripts that overflow cmd.exe after MSVC is already initialized.
+              mamba install -y --channel-priority flexible --override-channels -c $cudaChannel "${cudaChannel}::cuda-nvcc_win-64" "${cudaChannel}::cuda-cccl" "${cudaChannel}::cuda-libraries-dev=$cudaVersion" "${cudaChannel}::cuda-cudart" "${cudaChannel}::cuda-cudart-dev"
+            } else {
+              mamba install -y --channel-priority flexible --override-channels -c $cudaChannel "${cudaChannel}::cuda-toolkit=$cudaVersion" "${cudaChannel}::cuda-nvcc_win-64" "${cudaChannel}::cuda-cccl" "${cudaChannel}::cuda-cudart" "${cudaChannel}::cuda-cudart-dev"
+            }
+          } else {
+            throw 'Unsupported CUDA wheel build platform'
+          }
+          if ($LASTEXITCODE -ne 0) {
+            exit $LASTEXITCODE
+          }
+          if ($IsWindows) {
+            python -m pip install build wheel ninja
+          } else {
+            python -m pip install build wheel
+          }
+
+      - name: Build Wheel
+        run: |
+          $pathSeparator = if ($IsWindows) { ';' } else { ':' }
+          if ($IsWindows) {
+            $cudaRoot = Join-Path $env:CONDA_PREFIX 'Library'
+          } elseif (Test-Path (Join-Path $env:CONDA_PREFIX 'targets/x86_64-linux/include/cuda_runtime.h')) {
+            $cudaRoot = Join-Path $env:CONDA_PREFIX 'targets/x86_64-linux'
+          } else {
+            $cudaRoot = $env:CONDA_PREFIX
+          }
+
+          $env:CUDA_PATH = $cudaRoot
+          $env:CUDA_HOME = $cudaRoot
+          $env:CUDAToolkit_ROOT = $cudaRoot
+          $env:CUDA_TOOLKIT_ROOT_DIR = $cudaRoot
+          $cudaHostCompilerArg = ''
+          $cudaRootCmake = $cudaRoot.Replace('\', '/')
+          $env:CMAKE_ARGS = "-DCUDAToolkit_ROOT=$cudaRootCmake -DCUDA_TOOLKIT_ROOT_DIR=$cudaRootCmake"
+          if ($IsLinux) {
+            if ([version]$env:CUDAVER -lt [version]"12.0" -and (Test-Path '/usr/bin/g++-11')) {
+              $env:CC = '/usr/bin/gcc-11'
+              $env:CXX = '/usr/bin/g++-11'
+              $env:CUDAHOSTCXX = '/usr/bin/g++-11'
+              $cudaHostCompilerArg = " -DCMAKE_CUDA_HOST_COMPILER=$env:CUDAHOSTCXX"
+            } elseif (Test-Path '/usr/bin/g++-12') {
+              $env:CC = '/usr/bin/gcc-12'
+              $env:CXX = '/usr/bin/g++-12'
+              $env:CUDAHOSTCXX = '/usr/bin/g++-12'
+              $cudaHostCompilerArg = " -DCMAKE_CUDA_HOST_COMPILER=$env:CUDAHOSTCXX"
+            }
+            $env:CMAKE_ARGS = "-DCUDAToolkit_ROOT=$cudaRoot -DCUDA_TOOLKIT_ROOT_DIR=$cudaRoot$cudaHostCompilerArg"
+            $env:CPATH = "$cudaRoot/include$pathSeparator$env:CPATH"
+            $env:CPLUS_INCLUDE_PATH = "$cudaRoot/include$pathSeparator$env:CPLUS_INCLUDE_PATH"
+            $env:LIBRARY_PATH = "$cudaRoot/lib$pathSeparator$env:CONDA_PREFIX/lib$pathSeparator$env:LIBRARY_PATH"
+            $env:LD_LIBRARY_PATH = "$cudaRoot/lib$pathSeparator$env:CONDA_PREFIX/lib$pathSeparator$env:LD_LIBRARY_PATH"
+          } elseif ($IsWindows) {
+            $ninjaPath = ((Get-Command ninja -ErrorAction Stop).Source).Replace('\', '/')
+            $env:CMAKE_GENERATOR = 'Ninja'
+            $env:CMAKE_MAKE_PROGRAM = $ninjaPath
+            $env:PATH = "$(Join-Path $cudaRoot 'bin')$pathSeparator$env:PATH"
+          }
+
+          if ($IsWindows) {
+            $nvccCandidates = @(
+              (Join-Path $cudaRoot 'bin\nvcc.exe'),
+              (Join-Path $env:CONDA_PREFIX 'Library\bin\nvcc.exe'),
+              (Join-Path $env:CONDA_PREFIX 'bin\nvcc.exe')
+            )
+          } else {
+            $nvccCandidates = @(
+              (Join-Path $env:CONDA_PREFIX 'bin/nvcc'),
+              (Join-Path $env:CONDA_PREFIX 'targets/x86_64-linux/bin/nvcc')
+            )
+          }
+          $nvccPath = $nvccCandidates | Where-Object { Test-Path $_ } | Select-Object -First 1
+          if (-not $nvccPath) {
+            throw 'Failed to find nvcc in the conda environment'
+          }
+          $env:CUDACXX = $nvccPath
+          $env:PATH = "$(Split-Path $nvccPath)$pathSeparator$env:PATH"
+          if ($IsWindows) {
+            $nvccPathCmake = $nvccPath.Replace('\', '/')
+            $env:CUDACXX = $nvccPathCmake
+            $env:CMAKE_ARGS = "-DCMAKE_CUDA_COMPILER=$nvccPathCmake -DCMAKE_CUDA_COMPILER_ARG1=-allow-unsupported-compiler -DCMAKE_MAKE_PROGRAM=$env:CMAKE_MAKE_PROGRAM $env:CMAKE_ARGS"
+          }
+          $nvccVersion = ((& $nvccPath --version) | Select-String 'release ([0-9]+\.[0-9]+)').Matches[0].Groups[1].Value
+          if (-not $nvccVersion) {
+            throw 'Failed to detect the installed CUDA toolkit version'
+          }
+          $cudaTagVersion = $nvccVersion.Replace('.','')
+          $env:VERBOSE = '1'
+          $cudaArchs = "60-real;61-real;70-real;75-real;80-real;86-real;89-real;90-real;90-virtual"
+          if ([version]$nvccVersion -lt [version]"12.0") {
+            # CUDA 11.8 cannot compile ggml's Hopper PDL device calls.
+            $cudaArchs = "60-real;61-real;70-real;75-real;80-real;86-real;89-real"
+          } elseif ([version]$nvccVersion -ge [version]"13.0") {
+            # CUDA 13 dropped offline compilation support for pre-Turing targets.
+            $cudaArchs = "75-real;80-real;86-real;89-real;90-real;90-virtual"
+          }
+          # Build real cubins for the supported GPUs and keep one
+          # forward-compatible PTX target instead of embedding PTX for every SM.
+          $env:CMAKE_ARGS = "-DGGML_CUDA_FORCE_MMQ=ON -DGGML_CUDA=on -DCMAKE_CUDA_ARCHITECTURES=$cudaArchs -DCMAKE_CUDA_FLAGS=-allow-unsupported-compiler -DCMAKE_CUDA_FLAGS_INIT=-allow-unsupported-compiler $env:CMAKE_ARGS"
+          $env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DGGML_AVX2=off -DGGML_FMA=off -DGGML_F16C=off'
+          if ($env:AVXVER -eq 'basic') {
+            $env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DGGML_AVX=off'
+          }
+          python -m build --wheel
+          # Publish tags that reflect the actual installed toolkit version.
+          Write-Output "CUDA_VERSION=$cudaTagVersion" >> $env:GITHUB_ENV
+
+      - uses: softprops/action-gh-release@v2
+        if: startsWith(github.ref, 'refs/tags/')
+        with:
+          files: dist/*
+          # Set tag_name to <tag>-cu<cuda_version>.
+          tag_name: ${{ github.ref_name }}-cu${{ env.CUDA_VERSION }}
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

ggml_python-0.0.39/.github/workflows/wheels-index.yaml

@@ -0,0 +1,57 @@
+name: Wheels Index
+
+on:
+  # Trigger on any new release
+  release:
+    types: [published]
+
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+# Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
+permissions:
+  contents: read
+  pages: write
+  id-token: write
+
+# Allow only one concurrent deployment, skipping runs queued between the run in-progress and latest queued.
+# However, do NOT cancel in-progress runs as we want to allow these production deployments to complete.
+concurrency:
+  group: "pages"
+  cancel-in-progress: false
+
+jobs:
+  # Single deploy job since we're just deploying
+  deploy:
+    environment:
+      name: github-pages
+      url: ${{ steps.deployment.outputs.page_url }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Setup Pages
+        uses: actions/configure-pages@v5
+      - name: Build
+        run: |
+          ./scripts/releases-to-pep-503.sh index/whl/cpu '^[v]?[0-9]+\.[0-9]+\.[0-9]+$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu118 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu118$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu121 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu121$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu122 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu122$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu123 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu123$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu124 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu124$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu125 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu125$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu130 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu130$'
+          ./scripts/releases-to-pep-503.sh index/whl/cu132 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-cu132$'
+          ./scripts/releases-to-pep-503.sh index/whl/vulkan '^[v]?[0-9]+\.[0-9]+\.[0-9]+-vulkan$'
+          ./scripts/releases-to-pep-503.sh index/whl/rocm72 '^[v]?[0-9]+\.[0-9]+\.[0-9]+-rocm72$'
+          ./scripts/releases-to-pep-503.sh index/whl/hip-radeon '^[v]?[0-9]+\.[0-9]+\.[0-9]+-hip-radeon$'
+          ./scripts/releases-to-pep-503.sh index/whl/metal '^[v]?[0-9]+\.[0-9]+\.[0-9]+-metal$'
+      - name: Upload artifact
+        uses: actions/upload-pages-artifact@v3
+        with:
+          # Upload entire repository
+          path: "index"
+      - name: Deploy to GitHub Pages
+        id: deployment
+        uses: actions/deploy-pages@v4

ggml_python-0.0.39/.github/workflows/wheels-metal.yaml

@@ -0,0 +1,63 @@
+name: Wheels Metal
+
+on: workflow_dispatch
+
+permissions:
+  contents: write
+
+jobs:
+  build_wheels:
+    name: Build wheels on ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [macos-14, macos-15]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          submodules: "recursive"
+
+      # Used to host cibuildwheel.
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+          cache: "pip"
+
+      - name: Build wheels
+        uses: pypa/cibuildwheel@v2.22.0
+        env:
+          CIBW_REPAIR_WHEEL_COMMAND: ""
+          CIBW_ARCHS: "arm64"
+          CIBW_ENVIRONMENT: CMAKE_ARGS="-DCMAKE_OSX_ARCHITECTURES=arm64 -DCMAKE_APPLE_SILICON_PROCESSOR=arm64 -DGGML_NATIVE=off -DGGML_METAL=on -DCMAKE_CROSSCOMPILING=ON"
+          # The release wheel is tagged py3-none, so one build covers all
+          # supported Python versions and avoids duplicate wheel names.
+          CIBW_BUILD: "cp39-*"
+        with:
+          package-dir: .
+          output-dir: wheelhouse
+
+      - uses: actions/upload-artifact@v4
+        with:
+          name: wheels-mac_${{ matrix.os }}
+          path: ./wheelhouse/*.whl
+
+  release:
+    name: Release
+    needs: [build_wheels]
+    if: startsWith(github.ref, 'refs/tags/')
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/download-artifact@v4
+        with:
+          merge-multiple: true
+          path: dist
+
+      - uses: softprops/action-gh-release@v2
+        with:
+          files: dist/*
+          # Set release name to <tag>-metal.
+          tag_name: ${{ github.ref_name }}-metal
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}