Amazon SageMaker AI Features

You can run recipes within Amazon SageMaker HyperPod or as SageMaker training jobs. You use the HyperPod training adapter as the framework to help you run end-to-end training workflows. The training adapter is built on the NVIDIA NeMo framework and Neuronx Distributed Training package.

In Amazon SageMaker Studio, you can launch machine learning workloads on HyperPod clusters and view HyperPod cluster information. The increased visibility into cluster details and hardware metrics can help your team identify the right candidate for your pre-training or fine-tuning workloads.

Amazon SageMaker HyperPod task governance is a robust management system designed to streamline resource allocation and ensure efficient utilization of compute resources across teams and projects for your Amazon EKS clusters. HyperPod task governance also provides Amazon EKS cluster Observability, offering real-time visibility into cluster capacity, compute availability and usage, team allocation and utilization, and task run and wait time information.

With Amazon SageMaker Partner AI Apps, users get access to generative artificial intelligence (AI) and machine learning (ML) development applications built, published, and distributed by industry-leading application providers. Partner AI Apps are certified to run on SageMaker AI. With Partner AI Apps, users can accelerate and improve how they build solutions based on foundation models (FM) and classic ML models without compromising the security of their sensitive data, which stays completely within their trusted security configuration and is never shared with a third party.

You can chat with Amazon Q Developer in Amazon SageMaker Canvas using natural language for generative AI assistance with solving your machine learning problems. You can converse with Q Developer to discuss the steps of a machine learning workflow and leverage Canvas functionality such as data transforms, model building, and deployment.

Amazon SageMaker training plans are a compute reservation capability designed for large-scale AI model training workloads running on SageMaker training jobs and HyperPod clusters. They provide predictable access to high-demand GPU-accelerated computing resources within specified timelines. You can specify a desired timeline, duration, and maximum compute resources, and SageMaker training plans automatically manages infrastructure setup, workload execution, and fault recovery. This allows for efficiently planning and executing mission-critical AI projects with a predictable cost model.

An auto ML service that gives people with no coding experience the ability to build models and make predictions with them.

Code Editor extends Studio so that you can write, test, debug and run your analytics and machine learning code in an environment based on Visual Studio Code - Open Source ("Code-OSS").

Build, train, and deploy ML models using geospatial data.

Amazon SageMaker HyperPod is a capability of SageMaker AI that provides an always-on machine learning environment on resilient clusters that you can run any machine learning workloads for developing large machine learning models such as large language models (LLMs) and diffusion models.

JupyterLab in Studio improves latency and reliability for Studio Notebooks

Studio is the latest web-based experience for running ML workflows. Studio offers a suite of IDEs, including Code Editor, a new Jupyterlab application, RStudio, and Studio Classic.

An integrated machine learning environment where you can build, train, deploy, and analyze your models all in the same application.

A free service that gives customers access to AWS compute resources in an environment based on open-source JupyterLab.

An integrated development environment for R, with a console, syntax-highlighting editor that supports direct code execution, and tools for plotting, history, debugging and workspace management.

Build the workflows required for human review of ML predictions. Amazon A2I brings human review to all developers, removing the undifferentiated heavy lifting associated with building human review systems or managing large numbers of human reviewers.

Create an AutoML job to automatically train a model in Pipelines.

Users without machine learning knowledge can quickly build classification and regression models.

Preprocess datasets, run inference when you don't need a persistent endpoint, and associate input records with inferences to assist the interpretation of results.

Improve your machine learning models by detecting potential bias and help explain the predictions that models make.

A shared space consists of a shared JupyterServer application and a shared directory. All user profiles in a Amazon SageMaker AI domain have access to all shared spaces in the domain.

Import, analyze, prepare, and featurize data in SageMaker Studio. You can integrate Data Wrangler into your machine learning workflows to simplify and streamline data pre-processing and feature engineering using little to no coding. You can also add your own Python scripts and transformations to customize your data prep workflow.

Interact with your data, get visualizations, explore actionable insights, and fix data quality issues.

Inspect training parameters and data throughout the training process. Automatically detect and alert users to commonly occurring errors such as parameter values getting too large or small.

Optimize custom models for edge devices, create and manage fleets and run models with an efficient runtime.

Experiment management and tracking. You can use the tracked data to reconstruct an experiment, incrementally build on experiments conducted by peers, and trace model lineage for compliance and audit verifications.

A centralized store for features and associated metadata so features can be easily discovered and reused. You can create two types of stores, an Online or Offline store. The Online Store can be used for low latency, real-time inference use cases and the Offline Store can be used for training and batch inference.

High-quality training datasets by using workers along with machine learning to create labeled datasets.

A turnkey data labeling feature to create high-quality training datasets without having to build labeling applications and manage the labeling workforce on your own.

Get recommendations on inference instance types and configurations (e.g. instance count, container parameters and model optimizations) to use your ML models and workloads.

Evaluate any changes to your model-serving infrastructure by comparing its performance against the currently deployed infrastructure.

Learn about SageMaker AI features and capabilities through curated 1-click solutions, example notebooks, and pretrained models that you can deploy. You can also fine-tune the models and deploy them.

Track the lineage of machine learning workflows.

Create and manage machine learning pipelines integrated directly with SageMaker AI jobs.

Document information about your ML models in a single place for streamlined governance and reporting throughout the ML lifecycle.

A pre-built, visual overview of all the models in your account. Model Dashboard integrates information from SageMaker Model Monitor, transform jobs, endpoints, lineage tracking, and CloudWatch so you can access high-level model information and track model performance in one unified view.

Monitor and analyze models in production (endpoints) to detect data drift and deviations in model quality.

Versioning, artifact and lineage tracking, approval workflow, and cross account support for deployment of your machine learning models.

Train machine learning models once, then run anywhere in the cloud and at the edge.

Run your SageMaker Studio notebook as a non-interactive, scheduled job.

Analyze and preprocess data, tackle feature engineering, and evaluate models.

Create end-to-end ML solutions with CI/CD by using SageMaker Projects.

Maximize the long-term reward that an agent receives as a result of its actions.

Administrators can define least-privilege permissions for common ML activities using custom and preconfigured persona-based IAM roles.

A serverless endpoint option for hosting your ML model. Automatically scales in capacity to serve your endpoint traffic. Removes the need to select instance types or manage scaling policies on an endpoint.

A Git extension to enter the URL of a Git repository, clone it into your environment, push changes, and view commit history.

The next generation of SageMaker notebooks that include AWS IAM Identity Center (IAM Identity Center) integration, fast start-up times, and single-click sharing.

Easily discover, connect to, create, terminate and manage Amazon EMR clusters in single account and cross account configurations directly from SageMaker Studio.

Train deep learning models faster on scalable GPU instances managed by SageMaker AI.