I have successfully created the .mar file and a docker image, and pushed the image to AWS ECR. Afterwards I created a Sagemaker model using the ECR resource, and a Sagemaker endpoint using the model.
My question is, if I want to update the Sagemaker model (after local re-train, optimization, etc), is there a way to update it without pushing a new docker image onto ECR every time. Specifically, is there a way to convert the .mar file to a model.tar.gz file so that it can be used when creating a new/updated model?
Hi,
I have successfully created the .mar file and a docker image, and pushed the image to AWS ECR. Afterwards I created a Sagemaker model using the ECR resource, and a Sagemaker endpoint using the model.
My question is, if I want to update the Sagemaker model (after local re-train, optimization, etc), is there a way to update it without pushing a new docker image onto ECR every time. Specifically, is there a way to convert the .mar file to a model.tar.gz file so that it can be used when creating a new/updated model?
Thanks in advance!