Changes A new parameter called ExplainerConfig is added to CreateEndpointConfig API to enable SageMaker Clarify online explainability feature.
Changes SageMaker Training Managed Warm Pools let you retain provisioned infrastructure to reduce latency for repetitive training workloads.
Changes SageMaker now allows customization on Canvas Application settings, including enabling/disabling time-series forecasting and specifying an Amazon Forecast execution role at both the Domain and UserProfile levels.
Changes Amazon SageMaker Automatic Model Tuning now supports specifying Hyperband strategy for tuning jobs, which uses a multi-fidelity based tuning strategy to stop underperforming hyperparameter configurations early.
Changes This release adds Mode to AutoMLJobConfig.
Changes SageMaker Hosting now allows customization on ML instance storage volume size, model data download timeout and inference container startup ping health check timeout for each ProductionVariant in CreateEndpointConfig API.
Changes This release adds HyperParameterTuningJob type in Search API.
Changes This release enables administrators to attribute user activity and API calls from Studio notebooks, Data Wrangler and Canvas to specific users even when users share the same execution IAM role. ExecutionRoleIdentityConfig at Sagemaker domain level enables this feature.
Changes SageMaker Inference Recommender now accepts Inference Recommender fields: Domain, Task, Framework, SamplePayloadUrl, SupportedContentTypes, SupportedInstanceTypes, directly in our CreateInferenceRecommendationsJob API through ContainerConfig
Changes Amazon SageMaker Automatic Model Tuning now supports specifying multiple alternate EC2 instance types to make tuning jobs more robust when the preferred instance type is not available due to insufficient capacity.
Changes Amazon SageMaker Edge Manager provides lightweight model deployment feature to deploy machine learning models on requested devices.
Changes This release adds support for G5, P4d, and C6i instance types in Amazon SageMaker Inference and increases the number of hyperparameters that can be searched from 20 to 30 in Amazon SageMaker Automatic Model Tuning
Changes Heterogeneous clusters: the ability to launch training jobs with multiple instance types. This enables running component of the training job on the instance type that is most suitable for it. e.g. doing data processing and augmentation on CPU instances and neural network training on GPU instances
Changes This release adds: UpdateFeatureGroup, UpdateFeatureMetadata, DescribeFeatureMetadata APIs; FeatureMetadata type in Search API; LastModifiedTime, LastUpdateStatus, OnlineStoreTotalSizeBytes in DescribeFeatureGroup API.
Changes SageMaker Ground Truth now supports Virtual Private Cloud. Customers can launch labeling jobs and access to their private workforce in VPC mode.
Changes Amazon SageMaker Notebook Instances now allows configuration of Instance Metadata Service version and Amazon SageMaker Studio now supports G5 instance types.
Changes Amazon SageMaker Autopilot adds support for manually selecting features from the input dataset using the CreateAutoMLJob API.
Changes SageMaker Autopilot adds new metrics for all candidate models generated by Autopilot experiments; RStudio on SageMaker now allows users to bring your own development environment in a custom image.
Changes Amazon SageMaker Autopilot adds support for custom validation dataset and validation ratio through the CreateAutoMLJob and DescribeAutoMLJob APIs.
Changes SageMaker Inference Recommender now accepts customer KMS key ID for encryption of endpoints and compilation outputs created during inference recommendation.
Changes Amazon Sagemaker Notebook Instances now supports G5 instance types
Changes Autopilot now generates an additional report with information on the performance of the best model, such as a Confusion matrix and Area under the receiver operating characteristic (AUC-ROC). The path to the report can be found in CandidateArtifactLocations.
Changes This release added a new NNA accelerator compilation support for Sagemaker Neo.
Changes API changes relating to Fail steps in model building pipeline and add PipelineExecutionFailureReason in PipelineExecutionSummary.
Changes Amazon SageMaker now supports running training jobs on ml.g5 instance types.
Changes The release allows users to pass pipeline definitions as Amazon S3 locations and control the pipeline execution concurrency using ParallelismConfiguration. It also adds support of EMR jobs as pipeline steps.
Changes This release adds a new ContentType field in AutoMLChannel for SageMaker CreateAutoMLJob InputDataConfig.
Changes This release added a new Ambarella device(amba_cv2) compilation support for Sagemaker Neo.
Changes This release enables - 1/ Inference endpoint configuration recommendations and ability to run custom load tests to meet performance needs. 2/ Deploy serverless inference endpoints. 3/ Query, filter and retrieve end-to-end ML lineage graph, and incorporate model quality/bias detection in ML workflow.
Changes SageMaker CreateEndpoint and UpdateEndpoint APIs now support additional deployment configuration to manage traffic shifting options and automatic rollback monitoring. DescribeEndpoint now shows new in-progress deployment details with stage status.
Changes ListDevices and DescribeDevice now show Edge Manager agent version.
Changes This release adds support for RStudio on SageMaker.
Changes This release allows customers to describe one or more versioned model packages through BatchDescribeModelPackage, update project via UpdateProject, modify and read customer metadata properties using Create, Update and Describe ModelPackage and enables cross account registration of model packages.
Changes This release adds a new TrainingInputMode FastFile for SageMaker Training APIs.
Changes Add API for users to retry a failed pipeline execution or resume a stopped one.
Changes This release adds support for "Project Search"
Changes This release adds support for "Lifecycle Configurations" to SageMaker Studio
Changes Amazon SageMaker now supports Asynchronous Inference endpoints. Adds PlatformIdentifier field that allows Notebook Instance creation with different platform selections. Increases the maximum number of containers in multi-container endpoints to 15. Adds more instance types to InstanceType field.
Changes Amazon SageMaker Autopilot adds new metrics for all candidate models generated by Autopilot experiments.
Changes API changes with respect to Lambda steps in model building pipelines. Adds several waiters to async Sagemaker Image APIs. Add more instance types to AppInstanceType field
Changes Releasing new APIs related to Tuning steps in model building pipelines.
Changes SageMaker model registry now supports up to 5 containers and associated environment variables.
Changes Sagemaker Neo now supports running compilation jobs using customer's Amazon VPC
Changes Enable ml.g4dn instance types for SageMaker Batch Transform and SageMaker Processing
Changes Using SageMaker Edge Manager with AWS IoT Greengrass v2 simplifies accessing, maintaining, and deploying models to your devices. You can now create deployable IoT Greengrass components during edge packaging jobs. You can choose to create a device fleet with or without creating an AWS IoT role alias.
Changes AWS SageMaker - Releasing new APIs related to Callback steps in model building pipelines. Adds experiment integration to model building pipelines.
Changes Amazon SageMaker Autopilot now provides the ability to automatically deploy the best model to an endpoint
Changes Enable retrying Training and Tuning Jobs that fail with InternalServerError by setting RetryStrategy.
Changes Amazon SageMaker Autopilot now supports 1) feature importance reports for AutoML jobs and 2) PartialFailures for AutoML jobs
Changes This feature allows customer to specify the environment variables in their CreateTrainingJob requests.
Changes Adding authentication support for pulling images stored in private Docker registries to build containers for real-time inference.
Changes Support new target device ml_eia2 in SageMaker CreateCompilationJob API
Changes This release adds the ResolvedOutputS3Uri to the DescribeFeatureGroup API to indicate the S3 prefix where offline data is stored in a feature group
Changes Amazon SageMaker now supports core dump for SageMaker Endpoints and direct invocation of a single container in a SageMaker Endpoint that hosts multiple containers.
Changes This release adds expires-in-seconds parameter to the CreatePresignedDomainUrl API for support of a configurable TTL.
Changes Add a new optional FrameworkVersion field to Sagemaker Neo CreateCompilationJob API
Changes This feature allows customers to enable/disable model caching on Multi-Model endpoints.
Changes This feature helps you monitor model performance characteristics such as accuracy, identify undesired bias in your ML models, and explain model decisions better with explainability drift detection.
Changes Amazon SageMaker Pipelines for ML workflows. Amazon SageMaker Feature Store, a fully managed repository for ML features.
Changes This feature enables customers to encrypt their Amazon SageMaker Studio storage volumes with customer master keys (CMKs) managed by them in AWS Key Management Service (KMS).
Changes This release enables customers to bring custom images for use with SageMaker Studio notebooks.
Changes This release enables Sagemaker customers to convert Tensorflow and PyTorch models to CoreML (ML Model) format.
Changes This release adds support for launching Amazon SageMaker Studio in your VPC. Use AppNetworkAccessType in CreateDomain API to disable access to public internet and restrict the network traffic to VPC.
Changes Sagemaker Ground Truth: Added support for a new Streaming feature which helps to continuously feed data and receive labels in real time. This release adds a new input and output SNS data channel.
Changes Amazon SageMaker now supports 1) creating real-time inference endpoints using model container images from Docker registries in customers' VPC 2) AUC(Area under the curve) as AutoPilot objective metric
Changes Sagemaker Ground Truth:Added support for OIDC (OpenID Connect) to authenticate workers via their own identity provider instead of through Amazon Cognito. This release adds new APIs (CreateWorkforce, DeleteWorkforce, and ListWorkforces) to SageMaker Ground Truth service. Sagemaker Neo: Added support for detailed target device description by using TargetPlatform fields - OS, architecture, and accelerator. Added support for additional compilation parameters by using JSON field CompilerOptions. Sagemaker Search: SageMaker Search supports transform job details in trial components.
Changes This release adds the DeleteHumanTaskUi API to Amazon Augmented AI
Changes The new 'ModelClientConfig' parameter being added for CreateTransformJob and DescribeTransformJob api actions enable customers to configure model invocation related parameters such as timeout and retry.
Changes We are releasing HumanTaskUiArn as a new parameter in CreateLabelingJob and RenderUiTemplate which can take an ARN for a system managed UI to render a task.
Changes This release adds a new parameter (EnableInterContainerTrafficEncryption) to CreateProcessingJob API to allow for enabling inter-container traffic encryption on processing jobs.
Changes Change to the input, ResourceSpec, changing EnvironmentArn to SageMakerImageArn. This affects the following preview APIs: CreateDomain, DescribeDomain, UpdateDomain, CreateUserProfile, DescribeUserProfile, UpdateUserProfile, CreateApp and DescribeApp.
Changes Amazon SageMaker now supports running training jobs on ml.g4dn and ml.c5n instance types. Amazon SageMaker supports in "IN" operation for Search now.
Changes This release updates Amazon Augmented AI CreateFlowDefinition API and DescribeFlowDefinition response.
Changes SageMaker UpdateEndpoint API now supports retained variant properties, e.g., instance count, variant weight. SageMaker ListTrials API filter by TrialComponentName. Make ExperimentConfig name length limits consistent with CreateExperiment, CreateTrial, and CreateTrialComponent APIs.
Changes This release adds two new APIs (UpdateWorkforce and DescribeWorkforce) to SageMaker Ground Truth service for workforce IP whitelisting.
Changes SageMaker ListTrialComponents API filter by TrialName and ExperimentName.
Changes You can now use SageMaker Autopilot for automatically training and tuning candidate models using a combination of various feature engineering, ML algorithms, and hyperparameters determined from the user's input data. SageMaker Automatic Model Tuning now supports tuning across multiple algorithms. With Amazon SageMaker Experiments users can create Experiments, ExperimentTrials, and ExperimentTrialComponents to track, organize, and evaluate their ML training jobs. With Amazon SageMaker Debugger, users can easily debug training jobs using a number of pre-built rules provided by Amazon SageMaker, or build custom rules. With Amazon SageMaker Processing, users can run on-demand, distributed, and fully managed jobs for data pre- or post- processing or model evaluation. With Amazon SageMaker Model Monitor, a user can create MonitoringSchedules to automatically monitor endpoints to detect data drift and other issues and get alerted on them. This release also includes the preview version of Amazon SageMaker Studio with Domains, UserProfiles, and Apps. This release also includes the preview version of Amazon Augmented AI to easily implement human review of machine learning predictions by creating FlowDefinitions, HumanTaskUis, and HumanLoops.
Changes Amazon SageMaker now supports multi-model endpoints to host multiple models on an endpoint using a single inference container.
Changes Adds support for the new family of Elastic Inference Accelerators (eia2) for SageMaker Hosting and Notebook Services
Changes Enable G4D and R5 instances in SageMaker Hosting Services
Changes Amazon SageMaker now supports Amazon EFS and Amazon FSx for Lustre file systems as data sources for training machine learning models. Amazon SageMaker now supports running training jobs on ml.p3dn.24xlarge instance type. This instance type is offered as a limited private preview for certain SageMaker customers. If you are interested in joining the private preview, please reach out to the SageMaker Product Management team via AWS Support."
Changes Amazon SageMaker introduces Managed Spot Training. Increases the maximum number of metric definitions to 40 for SageMaker Training and Hyperparameter Tuning Jobs. SageMaker Neo adds support for Acer aiSage and Qualcomm QCS605 and QCS603.
Changes The default TaskTimeLimitInSeconds of labeling job is increased to 8 hours. Batch Transform introduces a new DataProcessing field which supports input and output filtering and data joining. Training job increases the max allowed input channels from 8 to 20.
Changes Workteams now supports notification configurations. Neo now supports Jetson Nano as a target device and NumberOfHumanWorkersPerDataObject is now included in the ListLabelingJobsForWorkteam response.
Changes Amazon SageMaker Automatic Model Tuning now supports random search and hyperparameter scaling.
Changes SageMaker notebook instances now support enabling or disabling root access for notebook users. SageMaker Neo now supports rk3399 and rk3288 as compilation target devices.
Changes SageMaker Training Jobs now support Inter-Container traffic encryption.
Changes Batch Transform Jobs now supports TFRecord as a Split Type. ListCompilationJobs API action now supports SortOrder and SortBy inputs.
Changes Amazon SageMaker Automatic Model Tuning now supports early stopping of training jobs. With early stopping, training jobs that are unlikely to generate good models will be automatically stopped during a Hyperparameter Tuning Job.
Changes Amazon SageMaker now has Algorithm and Model Package entities that can be used to create Training Jobs, Hyperparameter Tuning Jobs and hosted Models. Subscribed Marketplace products can be used on SageMaker to create Training Jobs, Hyperparameter Tuning Jobs and Models. Notebook Instances and Endpoints can leverage Elastic Inference accelerator types for on-demand GPU computing. Model optimizations can be performed with Compilation Jobs. Labeling Jobs can be created and supported by a Workforce. Models can now contain up to 5 containers allowing for inference pipelines within Endpoints. Code Repositories (such as Git) can be linked with SageMaker and loaded into Notebook Instances. Network isolation is now possible on Models, Training Jobs, and Hyperparameter Tuning Jobs, which restricts inbound/outbound network calls for the container. However, containers can talk to their peers in distributed training mode within the same security group. A Public Beta Search API was added that currently supports Training Jobs.
Changes SageMaker now makes the final set of metrics published from training jobs available in the DescribeTrainingJob results. Automatic Model Tuning now supports warm start of hyperparameter tuning jobs. Notebook instances now support a larger number of instance types to include instances from the ml.t3, ml.m5, ml.c4, ml.c5 families.
Changes SageMaker notebook instances can now have a volume size configured.
Changes VolumeKmsKeyId now available in Batch Transform Job
Changes Added an optional boolean parameter, 'DisassociateLifecycleConfig', to the UpdateNotebookInstance operation. When set to true, the lifecycle configuration associated with the notebook instance will be removed, allowing a new one to be set via a new 'LifecycleConfigName' parameter.