amazon-sagemaker-drift-detection

amazon-sagemaker-drift-detection copied to clipboard

Add support for batch pipeline that creates a SageMaker Pipeline that contains 3 steps 1. Source - pulls source code. 2. Build - get's the latest approved model, and outputs...

brightsparc

Add optional support to allow running build and deployment pipelines in GitHub actions, using [workflow approval](https://timheuer.com/blog/add-approval-workflow-to-github-actions) with the use of multiple environments and *protection rules*.

brightsparc

Replace the `evaluate.py` processing job with a [Lambda Step](https://docs.aws.amazon.com/sagemaker/latest/dg/build-and-manage-steps.html#step-type-lambda) that returns the report dictionary [Model Quality Metrics](https://docs.aws.amazon.com/sagemaker/latest/dg/model-monitor-model-quality-metrics.html) ``` # Get the training job response = sm_client.describe_training_job(TrainingJobName=job_name) status = response["TrainingJobStatus"] logger.info("Training...

brightsparc

Default AmazonSageMakerExecution Role missing permission to execute clean-up

2

The default AmazonSageMakerExecution is missing permissions to list and delete the stacks created by the SageMaker project. In particular, it is missing the following: ``` "cloudformation:DeleteStack" "cloudformation:DescribeStacks" ``` The default...

acere

brightsparc