CODEWITHSUNDEEP
amazon-web-servicesamazon-sagemaker
user4923
user4923

Reputation: 31

Store Model Evaluation Metrics to Sagemaker

currently I'm working in the setup of Sagemaker as our team ML Ops platform and we would like to store model evaluation metrics together with the model itself.

Now we are creating the model using the boto3 sagemaker client: https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/sagemaker.html#SageMaker.Client.create_model

But I don't see how to attach metrics to it, so we could run the model evaluation and store the metrics to keep track of them.

The goal is to have the model in Sagemaker so we can deploy it as an endpoint.

Thanks in advance!

Upvotes: 0

Views: 937

Answers (1)

Kirit Thadaka
Kirit Thadaka

Reputation: 507

You can attach metrics to the model if you register it in the Model Registry using the create_model_package API. There is an argument called ModelMetrics in the API that you can use for this. The metrics must be stored in a file in S3 and you will provide the S3 URI in this API call.

https://docs.aws.amazon.com/sagemaker/latest/APIReference/API_CreateModelPackage.html

Upvotes: 1

Related Questions