From 74a5e4e44cbb721fa1a646abe46a806bd96c4b9b Mon Sep 17 00:00:00 2001
From: ibpark-moreh <inbum.park@moreh.io>
Date: Wed, 4 Feb 2026 19:30:42 +0900
Subject: [PATCH] MAF-19213: Enhance Helm templates for DeepSeek and Qwen
 models

- Added model organization, name, role, accelerator vendor, model, and parallelism labels to DeepSeek-R1 decode and prefill templates.
- Updated quickstart templates for DeepSeek-R1 and Qwen models to include similar labels for consistency.
- Introduced role and parallelism labels in runtime base templates for decode and prefill configurations, ensuring better resource management and clarity across deployments.
---
 .../vllm-deepseek-r1-decode-mi300x-dp8ep.helm.yaml          | 6 ++++++
 .../vllm-deepseek-r1-prefill-mi300x-dp8ep.helm.yaml         | 6 ++++++
 ...epseek-ai-deepseek-r1-decode-amd-mi300x-dp8ep8.helm.yaml | 6 ++++++
 ...pseek-ai-deepseek-r1-prefill-amd-mi300x-dp8ep8.helm.yaml | 6 ++++++
 ...meta-llama-llama-3.2-1b-instruct-amd-mi250-tp2.helm.yaml | 6 ++++++
 ...eta-llama-llama-3.2-1b-instruct-amd-mi300x-tp2.helm.yaml | 6 ++++++
 ...ama-llama-3.2-1b-instruct-decode-amd-mi250-tp2.helm.yaml | 6 ++++++
 ...ma-llama-3.2-1b-instruct-decode-amd-mi300x-tp2.helm.yaml | 6 ++++++
 ...ma-llama-3.2-1b-instruct-prefill-amd-mi250-tp2.helm.yaml | 6 ++++++
 ...a-llama-3.2-1b-instruct-prefill-amd-mi300x-tp2.helm.yaml | 6 ++++++
 .../quickstart-vllm-qwen-qwen3-1.7b-amd-mi250-tp2.helm.yaml | 6 ++++++
 ...quickstart-vllm-qwen-qwen3-1.7b-amd-mi300x-tp2.helm.yaml | 6 ++++++
 ...tart-vllm-qwen-qwen3-1.7b-decode-amd-mi250-tp2.helm.yaml | 6 ++++++
 ...art-vllm-qwen-qwen3-1.7b-decode-amd-mi300x-tp2.helm.yaml | 6 ++++++
 ...art-vllm-qwen-qwen3-1.7b-prefill-amd-mi250-tp2.helm.yaml | 6 ++++++
 ...rt-vllm-qwen-qwen3-1.7b-prefill-amd-mi300x-tp2.helm.yaml | 6 ++++++
 .../templates/runtime-bases/vllm-decode-dp.helm.yaml        | 2 ++
 .../templates/runtime-bases/vllm-decode-pp.helm.yaml        | 2 ++
 .../templates/runtime-bases/vllm-decode.helm.yaml           | 1 +
 .../templates/runtime-bases/vllm-dp.helm.yaml               | 2 ++
 .../templates/runtime-bases/vllm-pp.helm.yaml               | 2 ++
 .../templates/runtime-bases/vllm-prefill-dp.helm.yaml       | 2 ++
 .../templates/runtime-bases/vllm-prefill-pp.helm.yaml       | 2 ++
 .../templates/runtime-bases/vllm-prefill.helm.yaml          | 1 +
 .../templates/runtime-bases/vllm.helm.yaml                  | 1 +
 25 files changed, 111 insertions(+)

diff --git a/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-decode-mi300x-dp8ep.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-decode-mi300x-dp8ep.helm.yaml
index b5d63c2..8310cb6 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-decode-mi300x-dp8ep.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-decode-mi300x-dp8ep.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: deepseek-ai
+    mif.moreh.io/model.name: deepseek-r1
+    mif.moreh.io/role: decode
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: dp8ep8
 spec:
   parallelism:
     data: 8
diff --git a/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-prefill-mi300x-dp8ep.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-prefill-mi300x-dp8ep.helm.yaml
index f321d08..2ba8d3d 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-prefill-mi300x-dp8ep.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/deepseek-r1/vllm-deepseek-r1-prefill-mi300x-dp8ep.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: deepseek-ai
+    mif.moreh.io/model.name: deepseek-r1
+    mif.moreh.io/role: prefill
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: dp8ep8
 spec:
   parallelism:
     data: 8
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-decode-amd-mi300x-dp8ep8.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-decode-amd-mi300x-dp8ep8.helm.yaml
index fa82ded..ddb5d9e 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-decode-amd-mi300x-dp8ep8.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-decode-amd-mi300x-dp8ep8.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: deepseek-ai
+    mif.moreh.io/model.name: deepseek-r1
+    mif.moreh.io/role: decode
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: dp8ep8
 spec:
   parallelism:
     data: 8
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-prefill-amd-mi300x-dp8ep8.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-prefill-amd-mi300x-dp8ep8.helm.yaml
index 556fc3d..db17433 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-prefill-amd-mi300x-dp8ep8.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-deepseek-ai-deepseek-r1-prefill-amd-mi300x-dp8ep8.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: deepseek-ai
+    mif.moreh.io/model.name: deepseek-r1
+    mif.moreh.io/role: prefill
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: dp8ep8
 spec:
   parallelism:
     data: 8
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi250-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi250-tp2.helm.yaml
index fe38209..9c8f3aa 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi250-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi250-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: meta-llama
+    mif.moreh.io/model.name: llama-3.2-1b-instruct
+    mif.moreh.io/role: e2e
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi250
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi300x-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi300x-tp2.helm.yaml
index 561ae37..2232afe 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi300x-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-amd-mi300x-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: meta-llama
+    mif.moreh.io/model.name: llama-3.2-1b-instruct
+    mif.moreh.io/role: e2e
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi250-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi250-tp2.helm.yaml
index 7f12bb6..8bdbd57 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi250-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi250-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: meta-llama
+    mif.moreh.io/model.name: llama-3.2-1b-instruct
+    mif.moreh.io/role: decode
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi250
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi300x-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi300x-tp2.helm.yaml
index 33b51c8..ba66c2f 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi300x-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-decode-amd-mi300x-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: meta-llama
+    mif.moreh.io/model.name: llama-3.2-1b-instruct
+    mif.moreh.io/role: decode
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi250-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi250-tp2.helm.yaml
index 4f34298..1e868a8 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi250-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi250-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: meta-llama
+    mif.moreh.io/model.name: llama-3.2-1b-instruct
+    mif.moreh.io/role: prefill
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi250
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi300x-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi300x-tp2.helm.yaml
index 946cc38..5d1b950 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi300x-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-meta-llama-llama-3.2-1b-instruct-prefill-amd-mi300x-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: meta-llama
+    mif.moreh.io/model.name: llama-3.2-1b-instruct
+    mif.moreh.io/role: prefill
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi250-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi250-tp2.helm.yaml
index 49a6b79..4394ace 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi250-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi250-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: qwen
+    mif.moreh.io/model.name: qwen3-1.7b
+    mif.moreh.io/role: e2e
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi250
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi300x-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi300x-tp2.helm.yaml
index 9cc538e..b9a4681 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi300x-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-amd-mi300x-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: qwen
+    mif.moreh.io/model.name: qwen3-1.7b
+    mif.moreh.io/role: e2e
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi250-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi250-tp2.helm.yaml
index 2fa2518..72c8af1 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi250-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi250-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: qwen
+    mif.moreh.io/model.name: qwen3-1.7b
+    mif.moreh.io/role: decode
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi250
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi300x-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi300x-tp2.helm.yaml
index 037bbfa..0d1a706 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi300x-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-decode-amd-mi300x-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: qwen
+    mif.moreh.io/model.name: qwen3-1.7b
+    mif.moreh.io/role: decode
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi250-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi250-tp2.helm.yaml
index 8b3ccb4..b9b14d6 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi250-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi250-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: qwen
+    mif.moreh.io/model.name: qwen3-1.7b
+    mif.moreh.io/role: prefill
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi250
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi300x-tp2.helm.yaml b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi300x-tp2.helm.yaml
index 936ed2f..c9ff80d 100644
--- a/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi300x-tp2.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/presets/quickstart/quickstart-vllm-qwen-qwen3-1.7b-prefill-amd-mi300x-tp2.helm.yaml
@@ -5,6 +5,12 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.preset.labels" . | nindent 4 }}
+    mif.moreh.io/model.org: qwen
+    mif.moreh.io/model.name: qwen3-1.7b
+    mif.moreh.io/role: prefill
+    mif.moreh.io/accelerator.vendor: amd
+    mif.moreh.io/accelerator.model: mi300x
+    mif.moreh.io/parallelism: tp2
 spec:
   parallelism:
     tensor: 2
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-dp.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-dp.helm.yaml
index 1bca467..8c6275b 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-dp.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-dp.helm.yaml
@@ -5,6 +5,8 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: decode
+    mif.moreh.io/parallelism: dp
 spec:
   workerTemplate:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-pp.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-pp.helm.yaml
index 9f8107d..a08a9b8 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-pp.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode-pp.helm.yaml
@@ -5,6 +5,8 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: decode
+    mif.moreh.io/parallelism: pp
 spec:
   template:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode.helm.yaml
index 94b1db1..553e4d0 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-decode.helm.yaml
@@ -5,6 +5,7 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: decode
 spec:
   template:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-dp.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-dp.helm.yaml
index 5ee02bd..e34bb12 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-dp.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-dp.helm.yaml
@@ -5,6 +5,8 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: e2e
+    mif.moreh.io/parallelism: dp
 spec:
   workerTemplate:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-pp.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-pp.helm.yaml
index 7928208..9219342 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-pp.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-pp.helm.yaml
@@ -5,6 +5,8 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: e2e
+    mif.moreh.io/parallelism: pp
 spec:
   template:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-dp.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-dp.helm.yaml
index 838596f..82e1935 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-dp.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-dp.helm.yaml
@@ -5,6 +5,8 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: prefill
+    mif.moreh.io/parallelism: dp
 spec:
   workerTemplate:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-pp.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-pp.helm.yaml
index 08cb6f2..7c36d42 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-pp.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill-pp.helm.yaml
@@ -5,6 +5,8 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: prefill
+    mif.moreh.io/parallelism: pp
 spec:
   template:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill.helm.yaml
index 578136c..ed0feb3 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm-prefill.helm.yaml
@@ -5,6 +5,7 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: prefill
 spec:
   template:
     metadata:
diff --git a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm.helm.yaml b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm.helm.yaml
index c6521ce..5be6118 100644
--- a/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm.helm.yaml
+++ b/deploy/helm/moai-inference-preset/templates/runtime-bases/vllm.helm.yaml
@@ -5,6 +5,7 @@ metadata:
   namespace: {{ include "common.names.namespace" . }}
   labels:
     {{- include "mif.runtimeBase.labels" . | nindent 4 }}
+    mif.moreh.io/role: e2e
 spec:
   template:
     spec: