pass args

clusters/k3s-dgx/nim-service/qwen.yaml

@@ -46,23 +46,15 @@ metadata:
   name: qwen36-27b-fp8
   namespace: nim-service
 spec:
-  args:
-    - "--gpu-memory-utilization=0.85"
-    - "--served-model-name=qwen36"
-    - "--max-model-len=256K"
-    - "--language-model-only"
-    - "--reasoning-parser=qwen3"
-    - "--enable-auto-tool-choice"
-    - "--tool-call-parser=qwen3_coder"
-    - "--enable-chunked-prefill"
-    - "--max-num-batched-tokens=32768"
-    - "--max-num-seqs=10"
-    - "--enable-prefix-caching"
-    - '--speculative-config={"method":"mtp","num_speculative_tokens":2}'
+  env:
+    - name: NIM_PASSTHROUGH_ARGS
+      value: "--gpu-memory-utilization 0.85 --served-model-name qwen36 --max-model-len 256K --language-model-only --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder --enable-chunked-prefill --max-num-batched-tokens 32768 --max-num-seqs 10 --enable-prefix-caching --speculative-config '{\"method\":\"mtp\",\"num_speculative_tokens\":2}'"
   image:
-    repository: vllm/vllm-openai
-    tag: "v0.19.1-cu130"
+    repository: nvcr.io/nim/nvidia/llm-nim
+    tag: "1.12"
     pullPolicy: IfNotPresent
+    pullSecrets:
+      - ngc-secret
   authSecret: hf-api-secret
   storage:
     nimCache: