hongbo-miao/hongbomiao.com

# https://github.com/pytorch/serve/blob/master/kubernetes/Helm/templates/torchserve.yaml

---
kind: Deployment
apiVersion: apps/v1
metadata:
  name: torchserve-deployment
  namespace: hm-cnn
  labels:
    app.kubernetes.io/name: torchserve
    app: torchserve
spec:
  replicas: 1
  selector:
    matchLabels:
      app: torchserve
  template:
    metadata:
      labels:
        app: torchserve
    spec:
      containers:
        - name: torchserve
          # image: docker.io/pytorch/torchserve:0.4.2-gpu
          image: docker.io/pytorch/torchserve:0.4.2-cpu
          args: ["torchserve", "--start", "--ts-config", "/data/model-server/config/config.properties"]
          ports:
            - name: inference
              protocol: TCP
              containerPort: 8080
            - name: inference-grpc
              protocol: TCP
              containerPort: 7070
            - name: management
              protocol: TCP
              containerPort: 8081
            - name: management-grpc
              protocol: TCP
              containerPort: 7071
            - name: metrics
              protocol: TCP
              containerPort: 8082
          volumeMounts:
            - name: model-server-volume
              mountPath: /data/model-server
          resources:
            limits:
              cpu: "1"
              memory: 4Gi
              nvidia.com/gpu: "0"
            requests:
              cpu: "1"
              memory: 1Gi
      volumes:
        - name: model-server-volume
          persistentVolumeClaim:
            claimName: model-server-pvc