fix(ci): pass CUDA compiler arg for Windows detection (#2202)

Author: abetlen

.github/workflows/build-wheels-cuda.yaml

@@ -156,7 +156,7 @@ jobs:
           if ($IsWindows) {
             $nvccPathCmake = $nvccPath.Replace('\', '/')
             $env:CUDACXX = $nvccPathCmake
-            $env:CMAKE_ARGS = "-DCMAKE_CUDA_COMPILER=$nvccPathCmake -DCMAKE_MAKE_PROGRAM=$env:CMAKE_MAKE_PROGRAM $env:CMAKE_ARGS"
+            $env:CMAKE_ARGS = "-DCMAKE_CUDA_COMPILER=$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) {
@@ -167,7 +167,7 @@ jobs:
           # Build real cubins for the supported GPUs, including sm_70, and keep
           # one forward-compatible PTX target instead of embedding PTX for every
           # SM. This keeps the wheel under GitHub's 2 GiB release-asset limit.
-          $env:CMAKE_ARGS = "-DGGML_CUDA_FORCE_MMQ=ON -DGGML_CUDA=on -DCMAKE_CUDA_ARCHITECTURES=70-real;75-real;80-real;86-real;89-real;90-real;90-virtual -DCMAKE_CUDA_FLAGS=--allow-unsupported-compiler -DCMAKE_CUDA_FLAGS_INIT=--allow-unsupported-compiler $env:CMAKE_ARGS"
+          $env:CMAKE_ARGS = "-DGGML_CUDA_FORCE_MMQ=ON -DGGML_CUDA=on -DCMAKE_CUDA_ARCHITECTURES=70-real;75-real;80-real;86-real;89-real;90-real;90-virtual -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'
           python -m build --wheel
           # Publish tags that reflect the actual installed toolkit version.