add ollama back to artemision

modules/kubernetes.nix

@@ -46,8 +46,33 @@
       kubernetes
     ];
 
-    # Enable containerd for Kubernetes
-    virtualisation.containerd.enable = true;
+    ## Enable containerd for Kubernetes
+    #virtualisation.containerd.enable = true;
 
+    ## Enable kubelet
+    #services.kubelet = {
+    #  enable = true;
+    #  extraFlags = {
+    #    "pod-infra-container-image" = "registry.k8s.io/pause:3.9";
+    #  };
+    #};
+
+    ## Enable kubeadm for cluster initialization
+    #environment.etc."kubeadm.yaml".text = ''
+    #  apiVersion: kubeadm.k8s.io/v1beta3
+    #  kind: InitConfiguration
+    #  localAPIEndpoint:
+    #    advertiseAddress: 127.0.0.1
+    #    bindPort: 6443
+    #  ---
+    #  apiVersion: kubeadm.k8s.io/v1beta3
+    #  kind: ClusterConfiguration
+    #  clusterName: ${config.services.kubernetes.clusterName}
+    #  controlPlaneEndpoint: ${config.services.kubernetes.controlPlaneEndpoint}
+    #  networking:
+    #    serviceSubnet: 10.96.0.0/12
+    #    podSubnet: 10.244.0.0/16
+    #    dnsDomain: cluster.local
+    #'';
   };
 }

systems/artemision/configuration.nix

@@ -1,6 +1,7 @@
 {
   lib,
   pkgs,
+  config,
   ...
 }:
 {
@@ -17,6 +18,7 @@
     ./stylix.nix
     ./wifi.nix
     ./zerotier.nix
+    ../palatine-hill/ollama.nix
   ];
 
   time.timeZone = "America/New_York";
@@ -39,6 +41,20 @@
   sops.age.sshKeyPaths = [ "/etc/ssh/ssh_host_ed25519_key" ];
 
   services = {
+    ollama = {
+      package = lib.mkForce pkgs.ollama-rocm;
+      models = lib.mkForce "${config.services.ollama.home}/models";
+      loadModels = lib.mkForce [
+        "deepseek-r1:1.5b"
+        "qwen3-coder-next"
+        "lennyerik/zeta"
+        "nomic-embed-text:latest"
+        "glm-4.7-flash"
+        "magistral"
+        "devstral-small-2"
+        "starcoder2:7b"
+      ];
+    };
     flatpak.enable = true;
     calibre-web = {
       # temp disable this

systems/palatine-hill/configuration.nix

@@ -17,7 +17,6 @@
     ./minio.nix
     ./networking.nix
     ./nextcloud.nix
-    ./ollama.nix
     #./plex
     ./postgresql.nix
     ./samba.nix

systems/palatine-hill/ollama.nix

@@ -25,13 +25,14 @@ in
         "nemotron-cascade-2:30b"
         "magistral"
         "devstral-small-2"
+        "starcoder2:15b"
       ];
       models = vars.primary_ollama;
       environmentVariables = {
         FLASH_ATTENTION = "1";
-        OLLAMA_KV_CACHE_TYPE = "q8_0";
+        OLLAMA_KV_CACHE_TYPE = "q4_0";
         # Ollama memory configuration
-        OLLAMA_MAX_LOADED_MODELS = "2";
+        OLLAMA_MAX_LOADED_MODELS = "3";
         OLLAMA_MAX_QUEUE = "512";
         OLLAMA_NUM_PARALLEL = "1";
 

users/alice/non-server.nix

@@ -85,7 +85,32 @@
             #keep_alive = "15m";
             available_models = [
               {
-                name = "glm-4.7-flash";
+                name = "deepseek-r1:1.5b";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "deepseek-r1:32b";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "deepseek-r1:70b";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "qwen3-coder-next";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "lennyerik/zeta";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "nomic-embed-text:latest";
                 max_tokens = 128000;
                 keep_alive = "15m";
               }
@@ -94,6 +119,22 @@
                 max_tokens = 128000;
                 keep_alive = "15m";
               }
+              {
+                name = "glm-4.7-flash";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "nemotron-cascade-2:30b";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+              {
+                name = "magistral";
+                max_tokens = 128000;
+                keep_alive = "15m";
+              }
+
             ];
           };
         };