Skip to Content

AWS launches new SageMaker features to make scaling machine learning easier

Posted on October 31, 2022 by

Categories: AWS

Tags:

SageMaker is Amazon Web Services (AWS) managed service for creating, training, and deploying machine learning (ML) models. The firm has released a plethora of new capabilities at todays re: Invent conference. Swami Sivasubramanian, Amazon’s VP of machine learning, has stated that the new capabilities are designed to facilitate the scalability of machine learning within businesses.

A new AWS service called SageMaker Ground Truth Plus leverages a skilled labor force to provide high-quality training datasets rapidly. Regarding labeling, SageMaker Ground Truth Plus employs a machine-learning approach that incorporates active learning, pre-labeling, and machine validation.

The new service, the business claims, can save expenses by as much as 40 percent and doesn’t need customers to have extensive knowledge of machine learning. Users can generate training datasets using the service without coding their labeling tools. Current availability of SageMaker Ground Truth Plus in the Northern Virginia area.

In addition, a new SageMaker Inference Recommender tool was released to aid customers in selecting the most suitable compute instance on which to install machine learning models in terms of performance and cost.

According to AWS, the tool can optimize your models and containers by choosing the best configurations for you. Except for the AWS China regions, Amazon SageMaker Inference Recommender usually is accessible in all locations where SageMaker is.

To further facilitate the deployment of machine learning models for inference without the need to install or manage the underlying infrastructure, AWS has launched a preview of a new SageMaker Serverless Interface option. This latest selection may be found in Northern Virginia, Ohio, Oregon, Ireland, Tokyo, and Sydney.

This morning, AWS introduced a new tool called SageMaker Training Compiler, which, by making better use of GPU instances, may speed up the training of deep learning models by as much as 50 percent.

Everything about deep learning models, from the abstract language they are written in down to the hardware-optimized instructions, is covered in this article. The newest function now lives in Northern Virginia, Ohio, Oregon, and Ireland.

Finally, Amazon Web Services (AWS) announced that customers may now see and debug Apache Spark operations running on Amazon Elastic MapReduce (EMR) from within SageMaker Studio notebooks.

SageMaker Studio can now be used to find and join EMR clusters, construct and delete nodes, and administer the entire infrastructure.

According to a blog post by Amazon Web Services, “the built-in connectivity with EMR enables you to execute interactive data preparation and machine learning at peta-byte scale directly within the one universal SageMaker Studio notebook.”

The newest SageMaker Studio enhancements are now accessible in the following locations: Northern Virginia, Ohio, Northern California, Oregon, central Canada, Frankfurt, Ireland, Stockholm, Ireland, Paris, London, Mumbai, Seoul, Singapore, Sydney, Tokyo, and Sao Paulo.

In a similar move, AWS today also introduced SageMaker Studio Lab, a no-cost service that educates and inspires developers to work with machine learning. Amazon Web Services (AWS) introduced Amazon SageMaker Canvas, a new machine learning service, yesterday. This new service will provide a simple interface for creating machine-learning prediction models.