ECS load container image from sagemaker built-in algorithm docker path

Question

Hello, I'd like to deploy SageMaker's built-in algorithm, BlazingText model on Fargate instead of Sagemaker endpoint. So, I tried to make an ECS task using BlazingText docker path. Here is my CDK code for it.

const loadBalancedFargateService = new ecsPatterns.ApplicationLoadBalancedFargateService(this, 'Service', {
            memoryLimitMiB: 1024,
            desiredCount: 1,
            cpu: 512,
            taskImageOptions: {
              image: ecs.ContainerImage.fromRegistry("811284229777.dkr.ecr.us-east-1.amazonaws.com/blazingtext:1"),
            },
          });

However, I got an error: 
CannotPullContainerError: inspect image has been retried 1 time(s): failed to resolve ref "811284229777.dkr.ecr.us-east-1.amazonaws.com/blazingtext:1": pulling from host 811284229777.dkr.ecr.us-east-1.amazonaws.com failed with status code [manifests 1]...

Is it impossible to pull docker container of sagemaker built-in algorithm from ECS?

Accepted Answer

To my knowledge, no - it's not generally possible to pull the [built-in algorithm](https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html) containers outside SageMaker: Your easiest route would probably just be to deploy the model on SageMaker and integrate your other containerized tasks to call the SageMaker endpoint.

It's maybe worth mentioning that the [framework containers](https://docs.aws.amazon.com/sagemaker/latest/dg/docker-containers-prebuilt.html) for custom/script-mode modelling (e.g. the [AWS DLCs](https://github.com/aws/deep-learning-containers) for PyTorch/HuggingFace/etc) are not subject to this restriction (can check you should even be able to pull them locally): So if you were to use those to implement a customized text processing model I think you should be able to deploy it on ECS if needed. Of course this'd mean a more initial build and later maintenance effort though.

ECS load container image from sagemaker built-in algorithm docker path

Relevant content