apiVersion: serving.kserve.io/v1alpha1
kind: ClusterServingRuntime
metadata:
  name: meta-nim-llama-3.3-70b-instruct
  namespace: default
spec:
  containers:
    - env:
        - name: NIM_CACHE_PATH
          value: /tmp
        - name: NGC_API_KEY
          valueFrom:
            secretKeyRef:
              name: nvidia-nim-secrets
              key: NGC_API_KEY
      image: upmdev.azurecr.io/nim/meta/llama-3.3-70b-instruct:1.8.4
      name: kserve-container
      ports:
        - containerPort: 8000
          protocol: TCP
      resources:
        limits:
          cpu: "24"
          memory: 600Gi
        requests:
          cpu: "24"
          memory: 600Gi
      volumeMounts:
        - mountPath: /dev/shm
          name: dshm
  imagePullSecrets:
    - name: edb-cred
  protocolVersions:
    - v2
    - grpc-v2
  supportedModelFormats:
    - autoSelect: true
      name: meta-nim-llama-3.3-70b-instruct
      priority: 1
      version: "1.8.4"
  volumes:
    - emptyDir:
        medium: Memory
        sizeLimit: 60Gi
      name: dshm