flake : update to support metal on m1/m2 (ggerganov#1724)

flake.nix

@@ -6,6 +6,13 @@
  outputs = { self, nixpkgs, flake-utils }:
  flake-utils.lib.eachDefaultSystem (system:
  let
+ inherit (pkgs.stdenv) isAarch64 isDarwin;
+ inherit (pkgs.lib) optionals;
+ isM1 = isAarch64 && isDarwin;
+ osSpecific =
+ if isM1 then with pkgs.darwin.apple_sdk_11_0.frameworks; [ Accelerate MetalKit MetalPerformanceShaders MetalPerformanceShadersGraph ]
+ else if isDarwin then with pkgs.darwin.apple_sdk.frameworks; [ Accelerate CoreGraphics CoreVideo ]
+ else [ ];
  pkgs = import nixpkgs {
  inherit system;
  };
@@ -18,17 +25,22 @@
  packages.default = pkgs.stdenv.mkDerivation {
  name = "llama.cpp";
  src = ./.;
+ postPatch =
+ if isM1 then ''
+ substituteInPlace ./ggml-metal.m \
+ --replace '[[NSBundle mainBundle] pathForResource:@"ggml-metal" ofType:@"metal"];' "@\"$out/ggml-metal.metal\";"
+ '' else "";
  nativeBuildInputs = with pkgs; [ cmake ];
- buildInputs = with pkgs; lib.optionals stdenv.isDarwin [
- darwin.apple_sdk.frameworks.Accelerate
- ];
- cmakeFlags = with pkgs; lib.optionals (system == "aarch64-darwin") [
+ buildInputs = osSpecific;
+ cmakeFlags = [ "-DLLAMA_BUILD_SERVER=ON" ] ++ (optionals isM1 [
  "-DCMAKE_C_FLAGS=-D__ARM_FEATURE_DOTPROD=1"
- ];
+ "-DLLAMA_METAL=ON"
+ ]);
  installPhase = ''
  mkdir -p $out/bin
  mv bin/* $out/bin/
  mv $out/bin/main $out/bin/llama
+ mv $out/bin/server $out/bin/llama-server
 
  echo "#!${llama-python}/bin/python" > $out/bin/convert.py
  cat ${./convert.py} >> $out/bin/convert.py
@@ -40,9 +52,7 @@
  packages = with pkgs; [
  cmake
  llama-python
- ] ++ lib.optionals stdenv.isDarwin [
- darwin.apple_sdk.frameworks.Accelerate
- ];
+ ] ++ osSpecific;
  };
  }
  );