Is it possible to deploy a scikit-learn model to a SageMaker serverless endpoint?

Question

I have trained a Scikit-Learn model within a SageMaker Training job and uploaded the model.tar.gz. file to S3. Now I would like to deploy this model as a serverless endpoint but I am getting the well-known error due to image size limitations of Lambda:
"Image size 10771903611 is greater than supported size 10737418240"
I am using image 492215442770.dkr.ecr.eu-central-1.amazonaws.com/sagemaker-scikit-learn:1.2-1-cpu-py3, whose size when pulled locally is 5.7GB (as output by `docker ls` command).
Is there an sklearn image suitable for serverless deployment?

Thanks,
Nikos

lavaraja-padala · Answer

Hello Nikos,

If possible please try to build the image from [official image](https://github.com/aws/sagemaker-scikit-learn-container/tree/master) and try with further reduced image size for deploying on serverless inference. You can exclude lines from  [L90-L116](https://github.com/aws/sagemaker-scikit-learn-container/blob/master/docker/1.2-1/base/Dockerfile.cpu#L90-L116) in the base image to reduce the size further and use the custom built final image to deploy the serverless inference endpoint. The steps for building the image are [here](https://github.com/aws/sagemaker-scikit-learn-container/blob/master/docker/1.2-1/base/Dockerfile.cpu#L90-L116).

Didier_Durand · Answer

Hi,

See part 5 of https://tutorialsdojo.com/train-and-deploy-a-scikit-learn-model-in-amazon-sagemaker/

It explains how to deploy a model trained with scikit-lean on AWS SageMaker

Best,

Didier

Is it possible to deploy a scikit-learn model to a SageMaker serverless endpoint?

Relevant content