Update example K8s manifests (#40)

kubernetes/data-volume.yaml

@@ -0,0 +1,28 @@
+# Create a PVC containing a model binary, sourced from an arbitrary HTTP server
+# (requires https://github.com/kubevirt/containerized-data-importer)
+apiVersion: cdi.kubevirt.io/v1beta1
+kind: DataVolume
+metadata:
+  name: models
+  namespace: local-ai
+spec:
+  contentType: archive
+  source:
+    http:
+      url: http://<model_server>/koala-7B-4bit-128g.GGML.tar
+      secretRef: model-secret
+  pvc:
+    accessModes:
+    - ReadWriteOnce
+    resources:
+      requests:
+        storage: 5Gi
+---
+apiVersion: v1
+kind: Secret
+metadata:
+  name: model-secret
+  namespace: local-ai
+data:
+  accessKeyId: <model_server_username_base64_encoded>
+  secretKey: <model_server_password_base64_encoded>

kubernetes/deployment.yaml

@@ -1,38 +1,55 @@
 apiVersion: v1
 kind: Namespace
 metadata:
-  name: llama
+  name: local-ai
 ---
 apiVersion: apps/v1
 kind: Deployment
 metadata:
-  name: llama
-  namespace: llama
+  name: local-ai
+  namespace: local-ai
   labels:
-    app: llama
+    app: local-ai
 spec:
   selector:
     matchLabels:
-      app: llama
+      app: local-ai
   replicas: 1
   template:
     metadata:
       labels:
-        app: llama
-      name: llama
+        app: local-ai
+      name: local-ai
     spec:
       containers:
-        - name: llama
+        - name: local-ai
           image: quay.io/go-skynet/local-ai:latest
+          env:
+          - name: THREADS
+            value: "14"
+          - name: CONTEXT_SIZE
+            value: "512"
+          - name: MODELS_PATH
+            value: /models
+          volumeMounts:
+          - mountPath: /models
+            name: models
+      volumes:
+      - name: models
+        persistentVolumeClaim:
+          claimName: models
 ---
 apiVersion: v1
 kind: Service
 metadata:
-  name: llama
-  namespace: llama
+  name: local-ai
+  namespace: local-ai
+  # If using AWS, you'll need to override the default 60s load balancer idle timeout
+  # annotations:
+  #   service.beta.kubernetes.io/aws-load-balancer-connection-idle-timeout: "1200"
 spec:
   selector:
-    app: llama
+    app: local-ai
   type: LoadBalancer
   ports:
     - protocol: TCP