Using R model in SageMaker ML pipelines

Question

Hi there,

Is it possible to use R model training and serving in SageMaker ML Pipelines? Looked in examples [here](https://github.com/aws/amazon-sagemaker-examples/tree/master/r_examples).  And it doesn't look that R is fully supported currently by ML Pipelines.  Any examples and success stories are very welcome.

Thanks.

Accepted Answer

In general it is possible to use the SageMaker python SDK and boto3  using the reticulate package in R, However do not have direct examples of SageMaker Pipelines using R.

It is possible to orchestrate the production pipeline using the R Containers for training and serving and setting up the DAG can be done with reticulate and SageMaker Python SDK and can be achieved using the AWS Step Functions. Please refer to the following example for reference.

https://github.com/aws-samples/reinvent2020-aim404-productionize-r-using-amazon-sagemaker
https://www.youtube.com/watch?v=Zpp0nfvqDCA

Using R model in SageMaker ML pipelines

Contenus pertinents