Merge pull request #142 from stackhpc/vllm-omni-2

Add vllm-omni backend support

Author: Alex-Welsh

charts/azimuth-chat/azimuth-ui.schema.yaml

@@ -24,6 +24,7 @@ sortOrder:
   - /azimuth-llm/api/azimuthNodeGroupSelector
   - /azimuth-llm/ui/appSettings/model_instruction
   - /azimuth-llm/ui/appSettings/page_title
+  - /azimuth-llm/api/image/containerImage
   - /azimuth-llm/api/image/version
   - /azimuth-llm/ui/appSettings/llm_params/temperature
   - /azimuth-llm/ui/appSettings/llm_params/max_tokens

charts/azimuth-chat/values.schema.json

@@ -36,10 +36,20 @@
                         "image": {
                             "type": "object",
                             "properties": {
+                                "containerImage": {
+                                    "type": "string",
+                                    "title": "vLLM Container Image",
+                                    "description": "Container to use as API backend. Currently only 'vllm/vllm-openai' and 'vllm/vllm-omni' are supported. Default is 'vllm/vllm-openai'.",
+                                    "default": "vllm/vllm-openai",
+                                    "enum": [
+                                        "vllm/vllm-openai",
+                                        "vllm/vllm-omni"
+                                    ]
+                                },
                                 "version": {
                                     "type": "string",
                                     "title": "Backend vLLM version",
-                                    "description": "The vLLM version to use as a backend. Must be a version tag from [this list](https://github.com/vllm-project/vllm/tags)",
+                                    "description": "The vLLM version to use as a backend. Must be a version tag from [this list](https://github.com/vllm-project/vllm/tags) when using vllm-openai or [this list](https://github.com/vllm-project/vllm-omni/tags) when using vllm-omni.",
                                     "default": "v0.11.0"
                                 }
                             }

charts/azimuth-image-analysis/azimuth-ui.schema.yaml

@@ -23,6 +23,7 @@ sortOrder:
   - /azimuth-llm/huggingface/token
   - /azimuth-llm/api/azimuthNodeGroupSelector
   - /azimuth-llm/ui/appSettings/page_title
+  - /azimuth-llm/api/image/containerImage
   - /azimuth-llm/api/image/version
   - /azimuth-llm/ui/appSettings/llm_params/temperature
   - /azimuth-llm/ui/appSettings/llm_params/max_tokens

charts/azimuth-image-analysis/values.schema.json

@@ -32,10 +32,20 @@
                         "image": {
                             "type": "object",
                             "properties": {
+                                "containerImage": {
+                                    "type": "string",
+                                    "title": "vLLM Container Image",
+                                    "description": "Container to use as API backend. Currently only 'vllm/vllm-openai' and 'vllm/vllm-omni' are supported. Default is 'vllm/vllm-openai'.",
+                                    "default": "vllm/vllm-openai",
+                                    "enum": [
+                                        "vllm/vllm-openai",
+                                        "vllm/vllm-omni"
+                                    ]
+                                },
                                 "version": {
                                     "type": "string",
                                     "title": "Backend vLLM version",
-                                    "description": "The vLLM version to use as a backend. Must be a version tag from [this list](https://github.com/vllm-project/vllm/tags)",
+                                    "description": "The vLLM version to use as a backend. Must be a version tag from [this list](https://github.com/vllm-project/vllm/tags) when using vllm-openai or [this list](https://github.com/vllm-project/vllm-omni/tags) when using vllm-omni.",
                                     "default": "v0.11.0"
                                 }
                             }

charts/azimuth-llm/azimuth-ui.schema.yaml

@@ -24,6 +24,7 @@ sortOrder:
   - /api/azimuthNodeGroupSelector
   - /ui/appSettings/model_instruction
   - /ui/appSettings/page_title
+  - /api/image/containerImage
   - /api/image/version
   - /ui/appSettings/llm_params/temperature
   - /ui/appSettings/llm_params/max_tokens

charts/azimuth-llm/templates/api/deployment.yml

@@ -24,7 +24,12 @@ spec:
         {{- else if .Values.api.intelXPUsEnabled }}
         image: "ghcr.io/stackhpc/vllm-xpu:{{ .Values.api.image.version }}"
         {{- else }}
-        image: "vllm/vllm-openai:{{ .Values.api.image.version }}"
+        image: "{{ .Values.api.image.containerImage }}:{{ .Values.api.image.version }}"
+        {{- end }}
+        {{- if eq .Values.api.image.containerImage "vllm/vllm-omni" }}
+        command:
+        - vllm
+        - serve
         {{- end }}
         ports:
         - name: api

charts/azimuth-llm/values.schema.json

@@ -33,10 +33,20 @@
                 "image": {
                     "type": "object",
                     "properties": {
+                        "containerImage": {
+                            "type": "string",
+                            "title": "vLLM Container Image",
+                            "description": "Container to use as API backend. Currently only 'vllm/vllm-openai' and 'vllm/vllm-omni' are supported. Default is 'vllm/vllm-openai'.",
+                            "default": "vllm/vllm-openai",
+                            "enum": [
+                                "vllm/vllm-openai",
+                                "vllm/vllm-omni"
+                            ]
+                        },
                         "version": {
                             "type": "string",
                             "title": "Backend vLLM version",
-                            "description": "The vLLM version to use as a backend. Must be a version tag from [this list](https://github.com/vllm-project/vllm/tags)",
+                            "description": "The vLLM version to use as a backend. Must be a version tag from [this list](https://github.com/vllm-project/vllm/tags) when using vllm-openai or [this list](https://github.com/vllm-project/vllm-omni/tags) when using vllm-omni.",
                             "default": "v0.11.0"
                         }
                     }

charts/azimuth-llm/values.yaml

@@ -37,6 +37,7 @@ api:
     # ghcr.io/stackhpc/vllm-xpu when api.gpus > 0 and intelXPUsEnabled is true,
     # or ghcr.io/stackhpc/vllm-cpu when api.gpus == 0
     repository:
+    containerImage: vllm/vllm-openai
     version: v0.11.0
   monitoring:
     enabled: true