apiVersion: sagemaker.aws.amazon.com/v1
kind: ProcessingJob
metadata:
  name: kmeans-mnist-processing
spec:
  environment:
    - name: MYVAR
      value: my_value
    - name: MYVAR2
      value: my_value2
  appSpecification:
    imageUri: 763104351884.dkr.ecr.us-west-2.amazonaws.com/pytorch-training:1.5.0-cpu-py36-ubuntu16.04
    containerEntrypoint:
      - python
      - /opt/ml/processing/code/kmeans_preprocessing.py
  roleArn: arn:aws:iam::123456789012:role/service-role/AmazonSageMaker-ExecutionRole
  region: us-west-2
  networkConfig:
    enableNetworkIsolation: False
  processingOutputConfig:
    outputs:
      - outputName: train_data
        s3Output:
          s3Uri: s3://my-bucket/mnist_kmeans_example/output/
          localPath: /opt/ml/processing/output_train/
          s3UploadMode: EndOfJob
      - outputName: test_data
        s3Output:
          s3Uri: s3://my-bucket/mnist_kmeans_example/output/
          localPath: /opt/ml/processing/output_test/
          s3UploadMode: EndOfJob
      - outputName: valid_data
        s3Output:
          s3Uri: s3://my-bucket/mnist_kmeans_example/output/
          localPath: /opt/ml/processing/output_valid/
          s3UploadMode: EndOfJob
  processingResources:
    clusterConfig:
      instanceCount: 1
      instanceType: ml.m5.xlarge
      volumeSizeInGB: 20
  processingInputs:
    - inputName: mnist_tar
      s3Input:
        s3Uri: s3://sagemaker-sample-data-us-west-2/algorithms/kmeans/mnist/mnist.pkl.gz
        localPath: /opt/ml/processing/input
        s3DataType: S3Prefix
        s3InputMode: File
        s3CompressionType: None
    - inputName: source_code
      s3Input:
        s3Uri: s3://my-bucket/mnist_kmeans_example/processing_code/kmeans_preprocessing.py
        localPath: /opt/ml/processing/code
        s3DataType: S3Prefix
        s3InputMode: File
        s3CompressionType: None
  stoppingCondition:
    maxRuntimeInSeconds: 1800
  tags:
    - key: tagKey
      value: tagValue