build intel mac with correct binary and compile flags

llm/llama.cpp/generate_darwin_amd64.go

@@ -14,6 +14,6 @@ package llm
 //go:generate git submodule update --force gguf
 //go:generate git -C gguf apply ../patches/0001-update-default-log-target.patch
 //go:generate git -C gguf apply ../patches/0001-metal-handle-ggml_scale-for-n-4-0-close-3754.patch
-//go:generate cmake -S gguf -B gguf/build/cpu -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=x86_64 -DCMAKE_OSX_ARCHITECTURES=x86_64 -DCMAKE_OSX_DEPLOYMENT_TARGET=11.0
+//go:generate cmake -S gguf -B gguf/build/cpu -DLLAMA_METAL=off -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=x86_64 -DCMAKE_OSX_ARCHITECTURES=x86_64 -DCMAKE_OSX_DEPLOYMENT_TARGET=11.0 -DLLAMA_NATIVE=off -DLLAMA_AVX=on -DLLAMA_AVX2=off -DLLAMA_AVX512=off -DLLAMA_FMA=off -DLLAMA_F16C=off
 //go:generate cmake --build gguf/build/cpu --target server --config Release
 //go:generate mv gguf/build/cpu/bin/server gguf/build/cpu/bin/ollama-runner

llm/llama.go

@@ -71,9 +71,10 @@ func chooseRunners(workDir, runnerType string) []ModelRunner {
 	// IMPORTANT: the order of the runners in the array is the priority order
 	switch runtime.GOOS {
 	case "darwin":
-		runners = []ModelRunner{
-			{Path: path.Join(buildPath, "metal", "bin", "ollama-runner")},
-			{Path: path.Join(buildPath, "cpu", "bin", "ollama-runner")},
+		if runtime.GOARCH == "arm64" {
+			runners = []ModelRunner{{Path: path.Join(buildPath, "metal", "bin", "ollama-runner")}}
+		} else {
+			runners = []ModelRunner{{Path: path.Join(buildPath, "cpu", "bin", "ollama-runner")}}
 		}
 	case "linux":
 		runners = []ModelRunner{

scripts/build_darwin.sh

@@ -10,6 +10,7 @@ mkdir -p dist
 for TARGETARCH in arm64 amd64; do
     GOOS=darwin GOARCH=$TARGETARCH go generate ./...
     GOOS=darwin GOARCH=$TARGETARCH go build -o dist/ollama-darwin-$TARGETARCH
+    rm -rf llm/llama.cpp/*/build
 done
 
 lipo -create -output dist/ollama dist/ollama-darwin-*