Sign inSign up

Samples Ml Mssql Image Classifier

Verified Publisher

samples/ml/mssql-image-classifier

By Microsoft

•Updated over 2 years ago

This is sample project that consumes an app deployed in SQL Server Big Data Cluster CTP 2.1.

0

microsoft/samples-ml-mssql-image-classifier repository overview

⁠Featured Tags

  • latest docker pull mcr.microsoft.com/samples/ml/mssql-image-classifier:latest

⁠About this Image

This is sample project to illustrate how you can consume an app deployed in SQL Server Big Data Cluster 2019 using CTP 2.1.

⁠How to Use this Image

To run it using docker:

# replace <cluster-ip>, <app-name>, <app-version> with actual values for your deployment of the Big Data Cluster and the app you created
docker run --rm -p 443:433 --env CLUSTER_IP="<cluster-ip>" --env APP_NAME="<app-name>" --env APP_VERSION="<app-version>" -it mcr.microsoft.com/samples/ml/mssql-image-classifier

Once it's running, point your browser to https://docker-host⁠ (replace docker-host with the your actual hostname or IP address of the host where you run the docker command)

To run it in Kubernetes, create a file mssql-image-classifier.yaml with the following content (replace <cluster-ip>, <app-name>, <app-version> with actual values):

apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: mssql-image-classifier
spec:
  replicas: 1
  template:
    metadata:
      labels:
        app: mssql-image-classifier
    spec:
      containers:
      - name: mssql-image-classifier
        image: mcr.microsoft.com/samples/ml/mssql-image-classifier
        env:
        - name: CLUSTER_IP
          value: <cluster-ip>
        - name: APP_NAME
          value: <app-name>
        - name: APP_VERSION
          value: <app-version>
        ports:
        - containerPort: 443
---
apiVersion: v1
kind: Service
metadata:
  name: mssql-image-classifier
spec:
  selector:
    app: mssql-image-classifier
  type: LoadBalancer
  ports:
  - protocol: TCP
    port: 443

And then run:

kubectl apply -f mssql-image-classifier.yaml

You can get the service's external endpoint (IP & port) by:

kubectl get svc mssql-image-classifier

If there isn't an external IP, use the node's IP and the node port of the mssql-image-classifier service from the output of the above command. The sample site is accessible through https.

To clean up:

kubectl delete svc mssql-image-classifier
kubectl delete deployment mssql-image-classifier

⁠Configuration

  • Requires Kubernetes cluster running SQL Server Big Data Cluster 2019 CTP 2.1 or newer, and compatible version of kubectl on a client machine.

⁠Full Tag Listing

You can retrieve a list of all available tags for samples/ml/mssql-image-classifier at https://mcr.microsoft.com/v2/samples/ml/mssql-image-classifier/tags/list⁠

⁠Support

For more information, go to https://support.microsoft.com/⁠

⁠License

Tag Summary

No tags have been pushed to this repository yet.