search

aws-samples / sagemaker-custom-project-templates

MIT No Attribution
214 stars 155 forks source link

Modify instance_count and instance_type to use Pipelines's Parameter … #75

Open hightensan opened 1 year ago

hightensan commented 1 year ago

…in batch-inference/seedcode/pipelines/batch_inference/pipeline.py

Issue #, if available: NA

Description of changes: In batch-inference/seedcode/pipelines/batch_inference/pipeline.py, instance_count and batch_inference_instance_type does not user pipeline's parameter. I replaced them with batch_inference_instance_count (ParameterInteger) and batch_inference_instance_type (ParameterString).

    #### PARAMETERS
    model_name = ParameterString("ModelName", default_value='${ModelName}')
    batch_inference_instance_count = ParameterInteger("BatchInstanceCount", default_value=1)
    batch_inference_instance_type = ParameterString("BatchInstanceType", default_value='ml.m5.xlarge')
    input_path = ParameterString("InputPath", default_value=f"s3://sagemaker-servicecatalog-seedcode-{region}/dataset/abalone-dataset.csv")
    output_path = ParameterString("OutputPath")

    #### SAGEMAKER CONSTRUCTS
    transform = Transformer(
        model_name=model_name,
        instance_count=1,
        instance_type='ml.m5.xlarge',
        output_path=output_path,
        base_transform_job_name=f"{base_job_prefix}/batch-transform-job",
        max_payload=10,
        accept='text/csv'
    )

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.