Add vllm chart in ai-stack (#18)

Signed-off-by: Sanket <[email protected]>

charts/ai-stack/Chart.lock

@@ -11,5 +11,8 @@ dependencies:
 - name: chromadb
   repository: https://infracloudio.github.io/charts
   version: 0.1.3
-digest: sha256:0febd220a71c6533c04a53affcfbeca2a77261acba6ded41f424cc34c2a056ff
-generated: "2024-08-19T19:46:03.544448+05:30"
+- name: vllm
+  repository: https://infracloudio.github.io/charts
+  version: 0.1.0
+digest: sha256:14b5e60e54b3618e5d950841fee42743eb9d50d2fed44d8d46484c97adbffde6
+generated: "2024-09-19T18:46:10.274968+05:30"

charts/ai-stack/Chart.yaml

@@ -16,7 +16,7 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.3.8
+version: 0.4.0
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to
@@ -49,6 +49,12 @@ dependencies:
     alias: vectordb
     condition: vectordb.enabled
 
+  - name: vllm
+    version: 0.1.0
+    repository: "https://infracloudio.github.io/charts"
+    alias: vllm
+    condition: vllm.enabled
+
 keywords:
   - ai-stack
   - ai-services

charts/ai-stack/README.md

@@ -13,6 +13,7 @@ The AI stack consists of the following components:
 - [Text Generation Inference(TGI)](../text-generation-inference/)
 - [Grafana Dashboards](../infracloud-dashboards)
 - [ChromaDB](../chromadb)
+- [vLLM](../vllm)
 
 ## Setup Helm Repository
 

charts/ai-stack/values.yaml

@@ -185,3 +185,49 @@ reranker:
     - name: hf-cache
       persistentVolumeClaim:
         claimName: hf-cache
+
+
+# Values for vllm: the vllm chart
+# Reference: https://artifacthub.io/packages/helm/infracloud-charts/vllm?modal=values
+vllm:
+  enabled: false
+
+  config:
+    model: "meta-llama/Meta-Llama-3.1-8B-Instruct"
+
+  env:
+  - name: HF_API_TOKEN
+    valueFrom:
+      secretKeyRef:
+        name: hf-api-token
+        key: HF_API_TOKEN
+  - name: HF_HUB_OFFLINE
+    value: "1"
+  - name: HF_HUB_CACHE
+    value: "/model"
+
+  resources:
+    limits:
+      nvidia.com/gpu: 1
+    requests:
+      nvidia.com/gpu: 1
+
+  strategy:
+    type: Recreate
+
+  service:
+    type: LoadBalancer
+    port: 8000
+
+  volumeMounts:
+    - name: hf-cache
+      mountPath: /model
+
+  volumes:
+    - name: hf-cache
+      persistentVolumeClaim:
+        claimName: hf-cache
+    - name: shm
+      emptyDir:
+        medium: Memory
+        sizeLimit: "1Gi"