Update test workflow

Author: jllllll

.github/workflows/build-wheels-test.yml

@@ -4,128 +4,53 @@ on:
   workflow_dispatch:
     inputs:
       version:
-        description: 'Version tag of llama-cpp-python to build: v0.2.2'
-        default: 'v0.2.2'
+        description: 'Version tag of llama-cpp-python to build: v0.2.6'
+        default: 'v0.2.6'
         required: false
         type: string
   workflow_call:
     inputs:
       version:
-        description: 'Version tag of llama-cpp-python to build: v0.2.2'
-        default: 'v0.2.2'
+        description: 'Version tag of llama-cpp-python to build: v0.2.6'
+        default: 'v0.2.6'
         required: false
         type: string
 
 jobs:
-  build_wheels:
-    name: Build Wheel ${{ matrix.os }} ${{ matrix.pyver }} ${{ matrix.cuda }} ${{ matrix.releasetag == 'wheels' && 'AVX2' || matrix.releasetag }}
+  build_wheels_main:
+    name: ${{ matrix.os }} ${{ matrix.pyver }} ${{ matrix.avx }}
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [ubuntu-20.04, windows-latest]
+        os: [macos-13, macos-12, macos-11]
+        arch: ["x86_64","arm64"]
         pyver: ["3.10"]
-        cuda: ["11.7.1"]
-        releasetag: ["wheels"]
-    defaults:
-      run:
-        shell: pwsh
     env:
-      CUDAVER: ${{ matrix.cuda }}
-      AVXVER: ${{ matrix.releasetag }}
+      CMAKE_OSX_ARCHITECTURES: ${{ matrix.arch }}
       PCKGVER: ${{ inputs.version }}
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
         with:
           repository: 'abetlen/llama-cpp-python'
           ref: ${{ inputs.version }}
           submodules: 'recursive'
 
-      - uses: actions/setup-python@v4
+      - uses: actions/setup-python@v3
         with:
           python-version: ${{ matrix.pyver }}
-        
-      - name: Setup Mamba
-        uses: conda-incubator/setup-miniconda@v2.2.0
-        with:
-          activate-environment: "build"
-          python-version: ${{ matrix.pyver }}
-          miniforge-variant: Mambaforge
-          miniforge-version: latest
-          use-mamba: true
-          add-pip-as-python-dependency: true
-          auto-activate-base: false
-          
-      - name: VS Integration Cache
-        id: vs-integration-cache
-        if: runner.os == 'Windows'
-        uses: actions/cache@v3.3.2
-        with:
-          path: ./MSBuildExtensions
-          key: cuda-${{ matrix.cuda }}-vs-integration
-          
-      - name: Get Visual Studio Integration
-        if: runner.os == 'Windows' && steps.vs-integration-cache.outputs.cache-hit != 'true'
-        run: |
-          if ($env:CUDAVER -eq '12.1.1') {$x = '12.1.0'} else {$x = $env:CUDAVER}
-          $links = (Invoke-RestMethod 'https://github.com/Jimver/cuda-toolkit/raw/257a101bc5c656053b5dc220126744980ef7f5b8/src/links/windows-links.ts').Trim().split().where({$_ -ne ''})
-          for ($i=$q=0;$i -lt $links.count -and $q -lt 2;$i++) {if ($links[$i] -eq "'$x',") {$q++}}
-          Invoke-RestMethod $links[$i].Trim("'") -OutFile 'cudainstaller.zip'
-          & 'C:\Program Files\7-Zip\7z.exe' e cudainstaller.zip -oMSBuildExtensions -r *\MSBuildExtensions\* > $null
-          Remove-Item 'cudainstaller.zip'
-          
-      - name: Install Visual Studio Integration
-        if: runner.os == 'Windows'
-        run: |
-          $y = (gi '.\MSBuildExtensions').fullname + '\*'
-          (gi 'C:\Program Files\Microsoft Visual Studio\2022\Enterprise\MSBuild\Microsoft\VC\*\BuildCustomizations').fullname.foreach({cp $y $_})
-          $cupath = 'CUDA_PATH_V' + $env:CUDAVER.Remove($env:CUDAVER.LastIndexOf('.')).Replace('.','_')
-          echo "$cupath=$env:CONDA_PREFIX" >> $env:GITHUB_ENV
 
       - name: Install Dependencies
         run: |
-          $cudaVersion = $env:CUDAVER
-          $cudaChannels = ''
-          $cudaNum = [int]$cudaVersion.substring($cudaVersion.LastIndexOf('.')+1)
-          while ($cudaNum -ge 0) { $cudaChannels += '-c nvidia/label/cuda-' + $cudaVersion.Remove($cudaVersion.LastIndexOf('.')+1) + $cudaNum + ' '; $cudaNum-- }
-          mamba install -y 'cuda' $cudaChannels.TrimEnd().Split()
-          if (!(mamba list cuda)[-1].contains('cuda')) {sleep -s 10; mamba install -y 'cuda' $cudaChannels.TrimEnd().Split()}
-          if (!(mamba list cuda)[-1].contains('cuda')) {throw 'CUDA Toolkit failed to install!'}
-          python -m pip install build wheel
+          python -m pip install build wheel cmake
           
       - name: Build Wheel
         run: |
-          $packageVersion = [version]$env:PCKGVER.TrimStart('v')
-          $cudaVersion = $env:CUDAVER.Remove($env:CUDAVER.LastIndexOf('.')).Replace('.','')
-          $env:CUDA_PATH = $env:CONDA_PREFIX
-          $env:CUDA_HOME = $env:CONDA_PREFIX
-          if ($IsLinux) {$env:LD_LIBRARY_PATH = $env:CONDA_PREFIX + '/lib:' + $env:LD_LIBRARY_PATH}
-          $env:VERBOSE = '1'
-          $env:CMAKE_ARGS = '-DLLAMA_CUBLAS=on -DCMAKE_CUDA_ARCHITECTURES=all'
-          if ($packageVersion -gt [version]'0.1.68' -and $packageVersion -lt [version]'0.1.71') {$env:CMAKE_ARGS = '-DLLAMA_CUBLAS=on -DCMAKE_CUDA_ARCHITECTURES=35-real;37-real;52;61-real;70-real;72-real;75-real;80-real;86-real;89-real;90'}
-          if ($packageVersion -gt [version]'0.1.68' -and $packageVersion -lt [version]'0.1.71' -and [version]$env:CUDAVER -ge [version]'12.0') {$env:CMAKE_ARGS = '-DLLAMA_CUBLAS=on -DCMAKE_CUDA_ARCHITECTURES=52;61-real;70-real;72-real;75-real;80-real;86-real;89-real;90'}
-          if ($packageVersion -gt [version]'0.1.68' -and $packageVersion -lt [version]'0.1.71' -and [version]$env:CUDAVER -lt [version]'11.8') {$env:CMAKE_ARGS = '-DLLAMA_CUBLAS=on -DCMAKE_CUDA_ARCHITECTURES=35-real;37-real;52;61-real;70-real;72-real;75-real;80-real;86'}
-          if ($packageVersion -lt [version]'0.1.66') {$env:CUDAFLAGS = '-arch=all'}
-          if ($env:AVXVER -eq 'AVX') {$env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DLLAMA_AVX2=off -DLLAMA_FMA=off -DLLAMA_F16C=off'}
-          if ($env:AVXVER -eq 'AVX512') {$env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DLLAMA_AVX512=on'}
-          if ($env:AVXVER -eq 'basic') {$env:CMAKE_ARGS = $env:CMAKE_ARGS + ' -DLLAMA_AVX=off -DLLAMA_AVX2=off -DLLAMA_FMA=off -DLLAMA_F16C=off'}
-          $buildtag = "+cu$cudaVersion"
-          if ($packageVersion -lt [version]'0.2.0') {
-            $env:FORCE_CMAKE = '1'
-            python -m build --wheel -C--build-option=egg_info "-C--build-option=--tag-build=$buildtag"
-          } else {
-            $initpath = Join-Path '.' 'llama_cpp' '__init__.py' -resolve
-            $initcontent = Get-Content $initpath -raw
-            $regexstr = '(?s)(?<=__version__ \= ")\d+(?:\.\d+)*(?=")'
-            $regexmatch = [Regex]::Matches($initcontent,$regexstr)
-            if (!($regexmatch[0].Success)) {throw '__init__.py parsing failed'}
-            $newinit = $regexmatch[0].Result(('$`' + '$&' + $buildtag + '$'''))
-            New-Item $initpath -itemType File -value $newinit -force
-            python -m build --wheel
-          }
+          export CMAKE_ARGS="-DLLAMA_METAL=on"
+          CMAKE_ARGS="-DLLAMA_METAL=on" VERBOSE=1 python -m build --wheel
         
       - uses: actions/upload-artifact@v3
         if: steps.upload-release.outcome == 'failure'
         with:
-          name: ${{ matrix.releasetag == 'wheels' && 'AVX2' || matrix.releasetag }}
+          name: ${{ format('wheel-{0}-{1}', matrix.os, matrix.arch)  }}
           path: ./dist/*.whl