Merge pull request #50 from defenseunicorns/helm-chart

Helm chart

chart/templates/api/configmap-model.yaml → chart/templates/api/configmap.yaml

@@ -1,7 +1,10 @@
 apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: api-model
 data:
-  models.toml: |
-{{ if .Values.models.stablelm3b.enabled }}
+  models.toml: | 
+{{- if .Values.models.stablelm3b.enabled }}
     [stablelm-3b]
         [stablelm-3b.metadata]
             owned_by    = 'Defense Unicorns'
@@ -11,12 +14,12 @@ data:
             url = 'stablelm:50051'
             type = 'gRPC'
 {{- end -}}
-{{- if .Values.models.all_MiniLM_L6_v2.enabled -}}
+{{ if .Values.models.all_MiniLM_L6_v2.enabled }}
     [text-embedding-ada-002]
         [text-embedding-ada-002.metadata]
             owned_by    = 'Defense Unicorns'
             permission  = []
-            description = 'A pretend implementation of ada-0002 tha's actually all-MiniLM-L6-v2'
+            description = 'A pretend implementation of ada-0002 thats actually all-MiniLM-L6-v2'
         [text-embedding-ada-002.network]
             url = 'sentence-transformers:50051'
             type = 'gRPC'   
@@ -27,9 +30,7 @@ data:
             description = ' By default, input text longer than 256 word pieces is truncated.'
         [all-MiniLM-L6-v2.network]
             url = 'sentence-transformers:50051'
-            type = 'gRPC'  
-{{- end -}} 
-kind: ConfigMap
-metadata:
-  creationTimestamp: null
-  name: api-model
+            type = 'gRPC'   
+{{- end -}}
+
+

chart/templates/api/deployment.yaml

@@ -17,4 +17,15 @@ spec:
           image: ghcr.io/defenseunicorns/leapfrogai/api:0.0.1
           imagePullPolicy: Always
           ports:
-            - containerPort: 50051
+            - containerPort: 50051
+          volumeMounts:
+          - name: api-model
+            mountPath: /app/models.toml
+            subPath: models.toml
+      volumes:
+      - name: api-model
+        configMap:
+          name: api-model
+
+
+

chart/templates/stablelm/deployment.yaml

@@ -0,0 +1,28 @@
+{{ if .Values.models.stablelm3b.enabled }}
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: stablelm-deployment
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: stablelm
+  template:
+    metadata:
+      labels:
+        app: stablelm
+    spec:
+      containers:
+        - name: stablelm-container
+          imagePullPolicy: Always
+          # imagePullPolicy: Never # Building on the node currently so we don't have to push/pull
+          image: ghcr.io/defenseunicorns/leapfrogai/stablelm-3b:0.0.1
+          ports:
+            - containerPort: 50051 # Assume this is the gRPC port
+          resources:
+            limits:
+              nvidia.com/gpu: 1
+            requests:
+              nvidia.com/gpu: 1
+{{ end }}

chart/templates/stablelm/service.yaml

@@ -0,0 +1,15 @@
+{{ if .Values.models.stablelm3b.enabled }}
+apiVersion: v1
+kind: Service
+metadata:
+  name: stablelm
+spec:
+  selector:
+    app: stablelm
+  ports:
+    - name: grpc
+      protocol: TCP
+      port: 50051
+      targetPort: 50051
+  type: ClusterIP
+{{ end }}

chart/templates/vs.yaml

@@ -7,7 +7,7 @@ spec:
   gateways:
   - istio-system/{{ .Values.istio.gateway }}
   hosts:
-  - leapfrogai.{{ .Values.domain}}
+  - leapfrogai.{{ .Values.domain }}
   http:
   - route:
     - destination:

chart/values.yaml

@@ -1,9 +1,9 @@
+domain: bigbang.dev
+
 istio:
   enabled: false
   gateway: public
 
-domain: bigbang.dev
-
 
 models:
   stablelm3b: