llama.cpp: things

services/llama-cpp.nix

@@ -9,7 +9,7 @@
 }:
 let
   cfg = config.services.llama-cpp;
-  modelUrl = "https://huggingface.co/bartowski/google_gemma-4-E2B-it-GGUF/resolve/main/google_gemma-4-E2B-it-Q4_K_M.gguf";
+  modelUrl = "https://huggingface.co/bartowski/google_gemma-4-E2B-it-GGUF/resolve/main/google_gemma-4-E2B-it-IQ2_M.gguf";
   modelAlias = lib.removeSuffix ".gguf" (baseNameOf modelUrl);
 in
 {
@@ -25,7 +25,7 @@ in
     model = toString (
       pkgs.fetchurl {
         url = modelUrl;
-        sha256 = "5efe645db4e1909c7a1f4a9608df18e6c14383f5e86777fc49f769f9ba7d5fdf";
+        sha256 = "17e869ac54d0e59faa884d5319fc55ad84cd866f50f0b3073fbb25accc875a23";
       }
     );
     port = service_configs.ports.private.llama_cpp.port;
@@ -33,7 +33,6 @@ in
     package = lib.optimizePackage (
       inputs.llamacpp.packages.${pkgs.system}.vulkan.overrideAttrs (old: {
         patches = (old.patches or [ ]) ++ [
-          ../patches/llamacpp/0003-gemma4-tokenizer-fix.patch
         ];
       })
     );