Awso_sagemaker.Values_0Sourceval structure_to_value_aux :
('a * 'b option) list ->
f:(('a * 'b) list -> 'c) ->
[> `Structure of 'c ]val structure_to_wrapped_value :
wrapper:'a ->
response:'a ->
('b * 'c option) list ->
[> `Structure of ('a * [> `Structure of ('b * 'c) list ]) list ]An inference component to benchmark.
The SageMaker endpoint configuration for benchmarking.
Summary information about an AI benchmark job.
Specifies an Amazon Virtual Private Cloud (VPC) that your SageMaker jobs, hosted models, and compute resources have access to. You can control access to and from your resources by configuring a VPC. For more information, see Give SageMaker Access to Resources in your Amazon VPC.
The network configuration for an AI benchmark job.
The output configuration for an AI benchmark job.
CloudWatch log information for an AI benchmark or recommendation job.
The output result of an AI benchmark job, including the Amazon S3 location and CloudWatch log information.
The target for an AI benchmark job. This is a union type — specify one of the members.
The capacity reservation configuration for an AI recommendation job.
The Amazon S3 data source for an AI workload.
The data source for an AI workload input data channel.
A channel of input data for an AI workload configuration. Each channel has a name and a data source.
The dataset configuration for an AI workload. This is a union type — specify one of the members.
The Amazon S3 model source configuration.
The source of the model for an AI recommendation job. This is a union type.
An expected performance metric for a recommendation.
Details about an optimization technique applied in a recommendation.
Instance details for a recommendation.
Details about the model package in a recommendation.
An Amazon S3 data channel for a recommended deployment configuration, containing model artifacts or optimized model outputs.
The deployment configuration for a recommendation.
An optimization recommendation generated by an AI recommendation job.
The compute resource specification for an AI recommendation job.
A performance constraint for an AI recommendation job.
The inference framework for an AI recommendation job.
Summary information about an AI recommendation job.
The output configuration for an AI recommendation job.
The output configuration for an AI recommendation job, including the S3 location for results and the model package group for deployment.
The performance targets for an AI recommendation job.
Summary information about an AI workload configuration.
The workload specification for benchmark tool configuration. Provide an inline YAML or JSON string.
The benchmark tool configuration for an AI workload.
Configuration for allocating accelerator partitions.
Configuration of the resources used for the compute allocation definition.
A structure describing the source of an action.
Lists the properties of an action. An action represents an action or activity. Some examples are a workflow step and a model deployment. Generally, an action involves at least one input artifact or output artifact.
Creates an association between the source and the destination. A source can be associated with multiple destinations, and a destination can be associated with multiple sources. An association is a lineage tracking entity. For more information, see Amazon SageMaker ML Lineage Tracking.
Resource being access is not found.
You have exceeded an SageMaker resource limit. For example, you might have too many training jobs created.
Creates an association between the source and the destination. A source can be associated with multiple destinations, and a destination can be associated with multiple sources. An association is a lineage tracking entity. For more information, see Amazon SageMaker ML Lineage Tracking.
Specifies an instance group and the number of nodes to add to it.
A tag object that consists of a key and an optional value, used to manage metadata for SageMaker Amazon Web Services resources. You can add tags to notebook instances, training jobs, hyperparameter tuning jobs, batch transform jobs, models, labeling jobs, work teams, endpoint configurations, and endpoints. For more information on adding tags to SageMaker resources, see AddTags. For more information on adding metadata to your Amazon Web Services resources with tagging, see Tagging Amazon Web Services resources. For advice on best practices for managing Amazon Web Services resources with tagging, see Tagging Best Practices: Implement an Effective Amazon Web Services Resource Tagging Strategy.
Adds or overwrites one or more tags for the specified SageMaker resource. You can add tags to notebook instances, training jobs, hyperparameter tuning jobs, batch transform jobs, models, labeling jobs, work teams, endpoint configurations, and endpoints. Each tag consists of a key and an optional value. Tag keys must be unique per resource. For more information about tags, see For more information, see Amazon Web Services Tagging Strategies. Tags that you add to a hyperparameter tuning job by calling this API are also added to any training jobs that the hyperparameter tuning job launches after you call this API, but not to training jobs that the hyperparameter tuning job launched before you called this API. To make sure that the tags associated with a hyperparameter tuning job are also added to all training jobs that the hyperparameter tuning job launches, add the tags when you first create the tuning job by specifying them in the Tags parameter of CreateHyperParameterTuningJob Tags that you add to a SageMaker Domain or User Profile by calling this API are also added to any Apps that the Domain or User Profile launches after you call this API, but not to Apps that the Domain or User Profile launched before you called this API. To make sure that the tags associated with a Domain or User Profile are also added to all Apps that the Domain or User Profile launches, add the tags when you first create the Domain or User Profile by specifying them in the Tags parameter of CreateDomain or CreateUserProfile.
Adds or overwrites one or more tags for the specified SageMaker resource. You can add tags to notebook instances, training jobs, hyperparameter tuning jobs, batch transform jobs, models, labeling jobs, work teams, endpoint configurations, and endpoints. Each tag consists of a key and an optional value. Tag keys must be unique per resource. For more information about tags, see For more information, see Amazon Web Services Tagging Strategies. Tags that you add to a hyperparameter tuning job by calling this API are also added to any training jobs that the hyperparameter tuning job launches after you call this API, but not to training jobs that the hyperparameter tuning job launched before you called this API. To make sure that the tags associated with a hyperparameter tuning job are also added to all training jobs that the hyperparameter tuning job launches, add the tags when you first create the tuning job by specifying them in the Tags parameter of CreateHyperParameterTuningJob Tags that you add to a SageMaker Domain or User Profile by calling this API are also added to any Apps that the Domain or User Profile launches after you call this API, but not to Apps that the Domain or User Profile launched before you called this API. To make sure that the tags associated with a Domain or User Profile are also added to all Apps that the Domain or User Profile launches, add the tags when you first create the Domain or User Profile by specifying them in the Tags parameter of CreateDomain or CreateUserProfile.
Information about additional Elastic Network Interfaces (ENIs) associated with an instance.
Input object for the model.
The access configuration file to control access to the ML model. You can explicitly accept the model end-user license agreement (EULA) within the ModelAccessConfig. If you are a Jumpstart user, see the End-user license agreements section for more details on accepting the EULA. If you are an AutoML user, see the Optional Parameters section of Create an AutoML job to fine-tune text generation models using the API for details on How to set the EULA acceptance when fine-tuning a model using the AutoML API.
Configuration information specifying which hub contents have accessible deployment options.
Specifies the S3 location of ML model data to deploy.
Specifies the location of ML model data to deploy. If specified, you must specify one and only one of the available data sources.
Identifies the foundation model that was used as the starting point for model customization.
A data source used for training or inference that is in addition to the input dataset or model data.
Data sources that are available to your model in addition to the one that you specify for ModelDataSource when you use the CreateModel action.
Describes the Docker container for the model package.
A structure of additional Inference Specification. Additional Inference Specification specifies details about inference jobs that can be run with models based on this model package
Edge Manager agent version.
An Amazon CloudWatch alarm configured to monitor metrics on an endpoint.
The details of the alarm to monitor during the AMI update.
An object containing authentication information for a private Docker registry.
The configuration to use an image from a private Docker registry for a training job.
Specifies a metric that the training algorithm writes to stderr or stdout. You can view these logs to understand how your training job performs and check for any errors encountered during training. SageMaker hyperparameter tuning captures all defined metrics. Specify one of the defined metrics to use as an objective metric using the TuningObjective parameter in the HyperParameterTrainingJobDefinition API to evaluate job performance during hyperparameter tuning.
Specifies the training algorithm to use in a CreateTrainingJob request. SageMaker uses its own SageMaker account credentials to pull and access built-in algorithms so built-in algorithms are universally accessible across all Amazon Web Services accounts. As a result, built-in algorithms have standard, unrestricted access. You cannot restrict built-in algorithms using IAM roles. Use custom algorithms if you require specific access controls. For more information about algorithms provided by SageMaker, see Algorithms. For information about using your own algorithms, see Using Your Own Algorithms with Amazon SageMaker.
Represents the overall status of an algorithm.
Specifies the validation and image scan statuses of the algorithm.
Provides summary information about an algorithm.
Describes the resources, including ML instance types and ML instance count, to use for transform job.
Describes the results of a transform job.
Describes the S3 data source.
Describes the location of the channel data.
Describes the input source of a transform job and the way the transform job consumes it.
Defines the input needed to run a transform job using the inference specification specified in the algorithm.
Maximum job scheduler pending time in seconds.
Specifies a limit to how long a job can run. When the job reaches the time limit, SageMaker ends the job. Use this API to cap costs. To stop a training job, SageMaker sends the algorithm the SIGTERM signal, which delays job termination for 120 seconds. Algorithms can use this 120-second window to save the model artifacts, so the results of training are not lost. The training algorithms provided by SageMaker automatically save the intermediate results of a model training job when possible. This attempt to save artifacts is only a best effort case as model might not be in a state from which it can be saved. For example, if training has just started, the model might not be ready to save. When saved, this intermediate data is a valid model artifact. You can use it to create a model with CreateModel. The Neural Topic Model (NTM) currently does not support saving intermediate model artifacts. When training NTMs, make sure that the maximum runtime is sufficient for the training job to complete.
Optional. Customer requested period in seconds for which the Training cluster is kept alive after the job is finished.
Specifies how instances should be placed on a specific UltraServer.
Configuration for how instances are placed and allocated within UltraServers. This is only applicable for UltraServer capacity.
Defines an instance group for heterogeneous cluster training. When requesting a training job using the CreateTrainingJob API, you can configure multiple instance groups .
Describes the resources, including machine learning (ML) compute instances and ML storage volumes, to use for model training.
Provides information about how to store model training results (model artifacts).
A configuration for a shuffle option for input data in a channel. If you use S3Prefix for S3DataType, the results of the S3 key prefix matches are shuffled. If you use ManifestFile, the order of the S3 object references in the ManifestFile is shuffled. If you use AugmentedManifestFile, the order of the JSON lines in the AugmentedManifestFile is shuffled. The shuffling order is determined using the Seed value. For Pipe input mode, when ShuffleConfig is specified shuffling is done at the start of every epoch. With large datasets, this ensures that the order of the training data is different for each epoch, and it helps reduce bias and possible overfitting. In a multi-node training job when ShuffleConfig is combined with S3DataDistributionType of ShardedByS3Key, the data is shuffled across nodes so that the content sent to a particular node on the first epoch might be sent to a different node on the second epoch.
The configuration for a private hub model reference that points to a public SageMaker JumpStart model. For more information about private hubs, see Private curated hubs for foundation model access control in JumpStart.
Describes the S3 data source. Your input bucket must be in the same Amazon Web Services region as your training job.
Specifies a file system data source for a channel.
Specifies a dataset source for a channel.
Describes the location of the channel data.
A channel is a named input source that training algorithms can consume.
Defines the input needed to run a training job using the algorithm.
Defines a training job and a batch transform job that SageMaker runs to validate your algorithm. The data provided in the validation profile is made available to your buyers on Amazon Web Services Marketplace.
Specifies configurations for one or more training jobs that SageMaker runs to test the algorithm.
A collection of settings that configure the Amazon Q experience within the domain.
Configures how labels are consolidated across human workers and processes output data.
TrainingPlanArn variant for ResourceSpec that allows "None" to detach a training plan. Based on TrainingPlanArn (min:50, max:2048) but with min:0 and "None" in pattern.
Specifies the ARN's of a SageMaker AI image and SageMaker AI image version, and the instance type that the version runs on. When both SageMakerImageVersionArn and SageMakerImageArn are passed, SageMakerImageVersionArn is used. Any updates to SageMakerImageArn will not take effect if SageMakerImageVersionArn already exists in the ResourceSpec because SageMakerImageVersionArn always takes precedence. To clear the value set for SageMakerImageVersionArn, pass None as the value.
Types duplicated from IronmanApiServiceModel for federation. These types are defined in other service directories and are not available via IronmanApiServiceCommonModel.
Details about an Amazon SageMaker AI app.
The specification of a Jupyter kernel.
The Amazon Elastic File System storage configuration for a SageMaker AI image.
The configuration for the file system and kernels in a SageMaker AI image running as a KernelGateway app.
The configuration used to run the application image container.
The configuration for the file system and kernels in a SageMaker AI image running as a JupyterLab app. The FileSystemConfig object is not supported.
The configuration for the file system and kernels in a SageMaker image running as a Code Editor app. The FileSystemConfig object is not supported.
The configuration for running a SageMaker AI image as a KernelGateway app.
Settings related to idle shutdown of Studio applications.
Settings that are used to configure and manage the lifecycle of Amazon SageMaker Studio applications.
Configuration to run a processing job in a specified container image.
The ID and ID type of an artifact source.
A structure describing the source of an artifact.
Lists a summary of the properties of an artifact. An artifact represents a URI addressable object or data. Some examples are a dataset and a model.
Associates a trial component with a trial. A trial component can be associated with multiple trials. To disassociate a trial component from a trial, call the DisassociateTrialComponent API.
Associates a trial component with a trial. A trial component can be associated with multiple trials. To disassociate a trial component from a trial, call the DisassociateTrialComponent API.
The data type used to describe the relationship between different sources.
The IAM Identity details associated with the user. These details are associated with model package groups, model packages and project entities only.
Information about the user who created or modified a SageMaker resource.
Lists a summary of the properties of an association. An association is an entity that links other lineage or experiment entities. An example would be an association between a training job and a model.
Configures the behavior of the client used by SageMaker to interact with the model container during asynchronous inference.
Specifies the configuration for notifications of inference results for asynchronous inference.
Specifies the configuration for asynchronous inference invocation outputs.
Specifies configuration for how an endpoint performs asynchronous inference.
The name of the data catalog used in Athena query execution.
The name of the database used in the Athena query execution.
The name of the workgroup in which the Athena query is being started.
The SQL query statements, to be executed.
Configuration for Athena Dataset Definition input.
Attaches your Amazon Elastic Block Store (Amazon EBS) volume to a node in your EKS orchestrated HyperPod cluster. This API works with the Amazon Elastic Block Store (Amazon EBS) Container Storage Interface (CSI) driver to manage the lifecycle of persistent storage in your HyperPod EKS clusters.
Attaches your Amazon Elastic Block Store (Amazon EBS) volume to a node in your EKS orchestrated HyperPod cluster. This API works with the Amazon Elastic Block Store (Amazon EBS) Container Storage Interface (CSI) driver to manage the lifecycle of persistent storage in your HyperPod EKS clusters.
Contains a presigned URL and its associated local file path for downloading hub content artifacts.
The selection of algorithms trained on your dataset to generate the model candidates for an Autopilot job.
The best candidate result from an AutoML training job.
Information about the steps for a candidate and what step it is working on.
Information about the metric for a candidate produced by an AutoML job.
The location of artifacts for an AutoML candidate job.
The properties of an AutoML candidate job.
A list of container definitions that describe the different containers that make up an AutoML candidate. For more information, see ContainerDefinition.
Information about a candidate produced by an AutoML training job, including its status, steps, and other properties.
Stores the configuration information for how a candidate is generated (optional).
Describes the Amazon S3 data source.
The data source for the Autopilot job.
A channel is a named input source that training algorithms can consume. The validation dataset size is limited to less than 2 GB. The training dataset size must be less than 100 GB. For more information, see Channel. A validation dataset must contain the same headers as the training dataset.
This data type is intended for use exclusively by SageMaker Canvas and cannot be used in other contexts at the moment. Specifies the compute configuration for the EMR Serverless job.
This data type is intended for use exclusively by SageMaker Canvas and cannot be used in other contexts at the moment. Specifies the compute configuration for an AutoML job V2.
This structure specifies how to split the data into train and validation datasets. The validation and training datasets must contain the same headers. For jobs created by calling CreateAutoMLJob, the validation dataset must be less than 2 GB in size.
The artifacts that are generated during an AutoML job.
A channel is a named input source that training algorithms can consume. This channel is used for AutoML jobs V2 (jobs created by calling CreateAutoMLJobV2).
How long a job is allowed to run, or how many candidates a job is allowed to generate.
Security options.
A collection of settings used for an AutoML job.
Specifies a metric to minimize or maximize as the objective of an AutoML job.
Metadata for an AutoML job step.
The reason for a partial failure of an AutoML job.
Provides a summary about an AutoML job.
The output data configuration.
Transformations allowed on the dataset. Supported transformations are Filling and Aggregation. Filling specifies how to add values to missing values in the dataset. Aggregation defines how to aggregate data that does not align with forecast frequency.
The collection of components that defines the time-series.
Stores the holiday featurization attributes applicable to each item of time-series datasets during the training of a forecasting model. This allows the model to identify patterns associated with specific holidays.
Stores the configuration information for how model candidates are generated using an AutoML job V2.
The collection of settings used by an AutoML job V2 for the time-series forecasting problem type.
The collection of settings used by an AutoML job V2 for the text generation problem type. The text generation models that support fine-tuning in Autopilot are currently accessible exclusively in regions supported by Canvas. Refer to the documentation of Canvas for the full list of its supported Regions.
The collection of settings used by an AutoML job V2 for the text classification problem type.
The collection of settings used by an AutoML job V2 for the tabular problem type.
The collection of settings used by an AutoML job V2 for the image classification problem type.
A collection of settings specific to the problem type used to configure an AutoML job V2. There must be one and only one config of the following type.
The resolved attributes specific to the text generation problem type.
The resolved attributes specific to the tabular problem type.
Stores resolved attributes specific to the problem type of an AutoML job V2.
The resolved attributes used to configure an AutoML job V2.
The name and an example value of the hyperparameter that you want to use in Autotune. If Automatic model tuning (AMT) determines that your hyperparameter is eligible for Autotune, an optimal hyperparameter range is selected for you.
Automatic rollback configuration for handling endpoint deployment failures and recovery.
A flag to indicate if you want to use Autotune to automatically find optimal values for the following fields: ParameterRanges: The names and ranges of parameters that a hyperparameter tuning job can optimize. ResourceLimits: The maximum resources that can be used for a training job. These resources include the maximum number of training jobs, the maximum runtime of a tuning job, and the maximum number of training jobs to run at the same time. TrainingJobEarlyStoppingType: A flag that specifies whether or not to use early stopping for training jobs launched by a hyperparameter tuning job. RetryStrategy: The number of times to retry a training job. Strategy: Specifies how hyperparameter tuning chooses the combinations of hyperparameter values to use for the training jobs that it launches. ConvergenceDetected: A flag to indicate that Automatic model tuning (AMT) has detected model convergence.
Contains information about an available upgrade for a SageMaker Partner AI App, including the version number and release notes.
Information about an error that occurred during the node addition operation.
Adds nodes to a HyperPod cluster by incrementing the target count for one or more instance groups. This operation returns a unique NodeLogicalId for each node being added, which can be used to track the provisioning status of the node. This API provides a safer alternative to UpdateCluster for scaling operations by avoiding unintended configuration changes. This API is only supported for clusters using Continuous as the NodeProvisioningMode.
Information about a node that was successfully added to the cluster.
Adds nodes to a HyperPod cluster by incrementing the target count for one or more instance groups. This operation returns a unique NodeLogicalId for each node being added, which can be used to track the provisioning status of the node. This API provides a safer alternative to UpdateCluster for scaling operations by avoiding unintended configuration changes. This API is only supported for clusters using Continuous as the NodeProvisioningMode.
Configuration to control how SageMaker captures inference data for batch transform jobs.
Information about an error that occurred when attempting to delete a node identified by its NodeLogicalId.
Represents an error encountered when deleting a node from a SageMaker HyperPod cluster.
Deletes specific nodes within a SageMaker HyperPod cluster. BatchDeleteClusterNodes accepts a cluster name and a list of node IDs. To safeguard your work, back up your data to Amazon S3 or an FSx for Lustre file system before invoking the API on a worker node group. This will help prevent any potential data loss from the instance root volume. For more information about backup, see Use the backup script provided by SageMaker HyperPod. If you want to invoke this API on an existing cluster, you'll first need to patch the cluster by running the UpdateClusterSoftware API. For more information about patching a cluster, see Update the SageMaker HyperPod platform software of a cluster.
Deletes specific nodes within a SageMaker HyperPod cluster. BatchDeleteClusterNodes accepts a cluster name and a list of node IDs. To safeguard your work, back up your data to Amazon S3 or an FSx for Lustre file system before invoking the API on a worker node group. This will help prevent any potential data loss from the instance root volume. For more information about backup, see Use the backup script provided by SageMaker HyperPod. If you want to invoke this API on an existing cluster, you'll first need to patch the cluster by running the UpdateClusterSoftware API. For more information about patching a cluster, see Update the SageMaker HyperPod platform software of a cluster.
The error code and error description associated with the resource.
This action batch describes a list of versioned model packages
Defines how to perform inference generation after a training job is run.
Provides summary information about the model package.
This action batch describes a list of versioned model packages
Represents an error encountered when rebooting a node (identified by its logical node ID) from a SageMaker HyperPod cluster.
Represents an error encountered when rebooting a node from a SageMaker HyperPod cluster.
Reboots specific nodes within a SageMaker HyperPod cluster using a soft recovery mechanism. BatchRebootClusterNodes performs a graceful reboot of the specified nodes by calling the Amazon Elastic Compute Cloud RebootInstances API, which attempts to cleanly shut down the operating system before restarting the instance. This operation is useful for recovering from transient issues or applying certain configuration changes that require a restart. Rebooting a node may cause temporary service interruption for workloads running on that node. Ensure your workloads can handle node restarts or use appropriate scheduling to minimize impact. You can reboot up to 25 nodes in a single request. For SageMaker HyperPod clusters using the Slurm workload manager, ensure rebooting nodes will not disrupt critical cluster operations.
Reboots specific nodes within a SageMaker HyperPod cluster using a soft recovery mechanism. BatchRebootClusterNodes performs a graceful reboot of the specified nodes by calling the Amazon Elastic Compute Cloud RebootInstances API, which attempts to cleanly shut down the operating system before restarting the instance. This operation is useful for recovering from transient issues or applying certain configuration changes that require a restart. Rebooting a node may cause temporary service interruption for workloads running on that node. Ensure your workloads can handle node restarts or use appropriate scheduling to minimize impact. You can reboot up to 25 nodes in a single request. For SageMaker HyperPod clusters using the Slurm workload manager, ensure rebooting nodes will not disrupt critical cluster operations.
Represents an error encountered when replacing a node (identified by its logical node ID) in a SageMaker HyperPod cluster.
Represents an error encountered when replacing a node in a SageMaker HyperPod cluster.
Replaces specific nodes within a SageMaker HyperPod cluster with new hardware. BatchReplaceClusterNodes terminates the specified instances and provisions new replacement instances with the same configuration but fresh hardware. The Amazon Machine Image (AMI) and instance configuration remain the same. This operation is useful for recovering from hardware failures or persistent issues that cannot be resolved through a reboot. Data Loss Warning: Replacing nodes destroys all instance volumes, including both root and secondary volumes. All data stored on these volumes will be permanently lost and cannot be recovered. To safeguard your work, back up your data to Amazon S3 or an FSx for Lustre file system before invoking the API on a worker node group. This will help prevent any potential data loss from the instance root volume. For more information about backup, see Use the backup script provided by SageMaker HyperPod. If you want to invoke this API on an existing cluster, you'll first need to patch the cluster by running the UpdateClusterSoftware API. For more information about patching a cluster, see Update the SageMaker HyperPod platform software of a cluster. You can replace up to 25 nodes in a single request.
Replaces specific nodes within a SageMaker HyperPod cluster with new hardware. BatchReplaceClusterNodes terminates the specified instances and provisions new replacement instances with the same configuration but fresh hardware. The Amazon Machine Image (AMI) and instance configuration remain the same. This operation is useful for recovering from hardware failures or persistent issues that cannot be resolved through a reboot. Data Loss Warning: Replacing nodes destroys all instance volumes, including both root and secondary volumes. All data stored on these volumes will be permanently lost and cannot be recovered. To safeguard your work, back up your data to Amazon S3 or an FSx for Lustre file system before invoking the API on a worker node group. This will help prevent any potential data loss from the instance root volume. For more information about backup, see Use the backup script provided by SageMaker HyperPod. If you want to invoke this API on an existing cluster, you'll first need to patch the cluster by running the UpdateClusterSoftware API. For more information about patching a cluster, see Update the SageMaker HyperPod platform software of a cluster. You can replace up to 25 nodes in a single request.
Represents the Parquet dataset format used when running a monitoring job.
Represents the JSON dataset format used when running a monitoring job.
Represents the CSV dataset format used when running a monitoring job.
Represents the dataset format used when running a monitoring job.
Input object for the batch transform job.
The metadata of the Amazon Bedrock custom model deployment.
The metadata of the Amazon Bedrock custom model.
The metadata of the Amazon Bedrock model import.
The metadata of the Amazon Bedrock provisioned model throughput.
A structure that keeps track of which training jobs launched by your hyperparameter tuning job are not improving model performance as evaluated against an objective function.
Details about the metrics source.
Specifies the type and size of the endpoint capacity to activate for a blue/green deployment, a rolling deployment, or a rollback strategy. You can specify your batches as either instance count or the overall percentage or your fleet. For a rollback strategy, if you don't specify the fields in this object, or if you set the Value to 100%, then SageMaker uses a blue/green rollback strategy and rolls all traffic back to the blue fleet.
Defines the traffic routing strategy during an endpoint deployment to shift traffic from the old fleet to the new fleet.
Update policy for a blue/green deployment. If this update policy is specified, SageMaker creates a new fleet during the deployment while maintaining the old fleet. SageMaker flips traffic to the new fleet according to the specified traffic routing configuration. Only one update policy should be used in the deployment configuration. If no update policy is specified, SageMaker uses a blue/green deployment strategy with all at once traffic shifting by default.
Details on the cache hit of a pipeline execution step.
An output parameter of a pipeline step.
Metadata about a callback step.
The workspace settings for the SageMaker Canvas application.
Time series forecast settings for the SageMaker Canvas application.
The model registry settings for the SageMaker Canvas application.
The Amazon SageMaker Canvas application setting where you configure document querying.
The Amazon SageMaker Canvas application setting where you configure OAuth for connecting to an external data source, such as Snowflake.
The generative AI settings for the SageMaker Canvas application. Configure these settings for Canvas users starting chats with generative AI foundation models. For more information, see Use generative AI with foundation models.
The settings for running Amazon EMR Serverless jobs in SageMaker Canvas.
The model deployment settings for the SageMaker Canvas application. In order to enable model deployment for Canvas, the SageMaker Domain's or user profile's Amazon Web Services IAM execution role must have the AmazonSageMakerCanvasDirectDeployAccess policy attached. You can also turn on model deployment permissions through the SageMaker Domain's or user profile's settings in the SageMaker console.
The SageMaker Canvas application settings.
Information about the Capacity Reservation used by an instance or instance group.
The configuration of the size measurements of the AMI update. Using this configuration, you can specify whether SageMaker should update your instance group by an amount or percentage of instances.
Configuration specifying how to treat different headers. If no headers are specified Amazon SageMaker AI will by default base64 encode when capturing the data.
Specifies data Model Monitor will capture.
Environment parameters you want to benchmark your load test against.
A list of categorical hyperparameters to tune.
Defines the possible values for a categorical hyperparameter.
A key-value pair that represents a parameter for the CloudFormation stack.
The CloudFormation template provider configuration for creating infrastructure resources.
Details about the CloudFormation stack.
A key-value pair representing a parameter used in the CloudFormation stack.
A key-value pair representing a parameter used in the CloudFormation stack.
Details about a CloudFormation template provider configuration and associated provisioning information.
Contains configuration details for updating an existing CloudFormation template provider in the project.
Defines a named input source, called a channel, to be used by an algorithm.
Contains information about the output location for managed spot training checkpoint data.
The container for the metadata for the ClarifyCheck step. For more information, see the topic on ClarifyCheck step in the Amazon SageMaker Developer Guide.
A parameter used to configure the SageMaker Clarify explainer to treat text features as text so that explanations are provided for individual units of text. Required only for natural language processing (NLP) explainability.
The configuration for the SHAP baseline (also called the background or reference dataset) of the Kernal SHAP algorithm. The number of records in the baseline data determines the size of the synthetic dataset, which has an impact on latency of explainability requests. For more information, see the Synthetic data of Configure and create an endpoint. ShapBaseline and ShapBaselineUri are mutually exclusive parameters. One or the either is required to configure a SHAP baseline.
The configuration for SHAP analysis using SageMaker Clarify Explainer.
The inference configuration parameter for the model container.
The configuration parameters for the SageMaker Clarify explainer.
Specifies the autoscaling configuration for a HyperPod cluster.
The autoscaling configuration and status information for a HyperPod cluster.
Configuration options specific to Spot instances.
Configuration options specific to On-Demand instances.
Defines the instance capacity requirements for an instance group, including configurations for both Spot and On-Demand capacity types.
Defines the configuration for attaching an additional Amazon Elastic Block Store (EBS) volume to each instance of the SageMaker HyperPod cluster instance group. To learn more, see SageMaker HyperPod release notes: June 20, 2024.
The customer ENI and additional ENIs associated with a network interface category.
Metadata information about an instance in a HyperPod cluster.
Metadata information about scaling operations for an instance group.
Metadata information about an instance group in a HyperPod cluster.
Metadata information about a HyperPod cluster showing information about the cluster level operations, such as creating, updating, and deleting.
Metadata associated with a cluster event, which may include details about various resource types.
Detailed information about a specific event, including event metadata.
Detailed information about a specific event in a HyperPod cluster.
A summary of an event in a HyperPod cluster.
Defines the configuration for attaching an Amazon FSx for Lustre file system to instances in a SageMaker HyperPod cluster instance group.
Defines the configuration for attaching an Amazon FSx for OpenZFS file system to instances in a SageMaker HyperPod cluster instance group.
The configurations that SageMaker uses when updating the AMI versions.
The configuration to use when updating the AMI versions.
The configuration object of the schedule that SageMaker follows when updating the AMI.
The Slurm configuration details for an instance group in a SageMaker HyperPod cluster.
The network interface configuration details for a Amazon SageMaker HyperPod cluster instance group.
The lifecycle configuration for a SageMaker HyperPod cluster.
A Kubernetes taint that can be applied to cluster nodes.
Detailed Kubernetes configuration showing both the current and desired state of labels and taints for cluster nodes.
Details about a specific instance type within a flexible instance group, including the count and configuration.
Defines the configuration for attaching additional storage to the instances in the SageMaker HyperPod cluster instance group. To learn more, see SageMaker HyperPod release notes: June 20, 2024.
The instance requirement details for a flexible instance group, including the current and desired instance types.
Details of an instance group in a SageMaker HyperPod cluster.
The Slurm configuration for an instance group in a SageMaker HyperPod cluster.
The network interface configuration for a Amazon SageMaker HyperPod cluster instance group.
Kubernetes configuration that specifies labels and taints to be applied to cluster nodes in an instance group.
The instance requirements for a flexible instance group. Use this to specify multiple instance types that the instance group can use. The order of instance types in the list determines the priority for instance provisioning.
The specifications of an instance group that you need to define.
Specifies the placement details for the node in the SageMaker HyperPod cluster, including the Availability Zone and the unique identifier (ID) of the Availability Zone.
Details of an instance in a SageMaker HyperPod cluster.
Node-specific Kubernetes configuration showing both current and desired state of labels and taints for an individual cluster node.
Contains information about the UltraServer object.
Details of an instance (also called a node interchangeably) in a SageMaker HyperPod cluster.
Lists a summary of the properties of an instance (also called a node interchangeably) of a SageMaker HyperPod cluster.
The configuration settings for the Slurm orchestrator used with the SageMaker HyperPod cluster.
The configuration settings for the Amazon EKS cluster used as the orchestrator for the SageMaker HyperPod cluster.
The type of orchestrator used for the SageMaker HyperPod cluster.
Configuration settings for an Amazon FSx for Lustre file system to be used with the cluster.
The configuration details for the restricted instance groups (RIG) environment.
The instance group details of the restricted instance group (RIG).
The configuration for the restricted instance groups (RIG) environment.
The specifications of a restricted instance group that you need to define.
Summary of the cluster policy.
Lists a summary of the properties of a SageMaker HyperPod cluster.
Defines the configuration for managed tier checkpointing in a HyperPod cluster. Managed tier checkpointing uses multiple storage tiers, including cluster CPU memory, to provide faster checkpoint operations and improved fault tolerance for large-scale model training. The system automatically saves checkpoints at high frequency to memory and periodically persists them to durable storage, like Amazon S3.
A custom SageMaker AI image. For more information, see Bring your own SageMaker AI image.
The Code Editor application settings. For more information about Code Editor, see Get started with Code Editor in Amazon SageMaker.
A Git repository that SageMaker AI automatically displays to users for cloning in the JupyterServer application.
Specifies configuration details for a Git repository in your Amazon Web Services account.
Specifies summary information about a Git repository.
Use this parameter to configure your Amazon Cognito workforce. A single Cognito workforce is created using and corresponds to a single Amazon Cognito user pool.
Identifies a Amazon Cognito user group. A user group can be used in on or more work teams.
Configuration for your vector collection type.
Configuration for your collection.
Configuration information for the Amazon SageMaker Debugger output tensor collections.
A summary of a model compilation job.
Resource sharing configuration.
Configuration of the compute allocation definition for an entity. This includes the resource sharing option and the setting to preempt low priority tasks.
The target entity to allocate compute resources to.
Summary of the compute allocation definition.
Metadata for a Condition step.
There was a conflict when you attempted to modify a SageMaker entity such as an Experiment or Artifact.
Specifies additional configuration for hosting multi-model endpoints.
Specifies an authentication configuration for the private docker registry where your model image is hosted. Specify a value for this property only if you specified Vpc as the value for the RepositoryAccessMode field of the ImageConfig object that you passed to a call to CreateModel and the private Docker registry where the model image is hosted requires authentication.
Specifies whether the model container is in Amazon ECR or a private Docker registry accessible from your Amazon Virtual Private Cloud (VPC).
Describes the container, as part of model definition.
A structure describing the source of a context.
Lists a summary of the properties of a context. A context provides a logical grouping of other entities.
A list of continuous hyperparameters to tune.
Defines the possible values for a continuous hyperparameter.
A flag to indicating that automatic model tuning (AMT) has detected model convergence, defined as a lack of significant improvement (1% or less) against an objective metric.
Creates a benchmark job that runs performance benchmarks against inference infrastructure using a predefined AI workload configuration. The benchmark job measures metrics such as latency, throughput, and cost for your generative AI inference endpoints.
Resource being accessed is in use.
Creates a benchmark job that runs performance benchmarks against inference infrastructure using a predefined AI workload configuration. The benchmark job measures metrics such as latency, throughput, and cost for your generative AI inference endpoints.
Creates a recommendation job that generates intelligent optimization recommendations for generative AI inference deployments. The job analyzes your model, workload configuration, and performance targets to recommend optimal instance types, model optimization techniques (such as quantization and speculative decoding), and deployment configurations.
Creates a recommendation job that generates intelligent optimization recommendations for generative AI inference deployments. The job analyzes your model, workload configuration, and performance targets to recommend optimal instance types, model optimization techniques (such as quantization and speculative decoding), and deployment configurations.
Creates a reusable AI workload configuration that defines datasets, data sources, and benchmark tool settings for consistent performance testing of generative AI inference deployments on Amazon SageMaker AI.
Creates a reusable AI workload configuration that defines datasets, data sources, and benchmark tool settings for consistent performance testing of generative AI inference deployments on Amazon SageMaker AI.
Metadata properties of the tracking entity, trial, or trial component.
Creates an action. An action is a lineage tracking entity that represents an action or activity. For example, a model deployment or an HPO job. Generally, an action involves at least one input or output artifact. For more information, see Amazon SageMaker ML Lineage Tracking.
Creates an action. An action is a lineage tracking entity that represents an action or activity. For example, a model deployment or an HPO job. Generally, an action involves at least one input or output artifact. For more information, see Amazon SageMaker ML Lineage Tracking.
Defines the objective metric for a hyperparameter tuning job. Hyperparameter tuning uses the value of this metric to evaluate the training jobs it launches, and returns the training job that results in either the highest or lowest value for this metric, depending on the value you specify for the Type parameter. If you want to define a custom objective metric, see Define metrics and environment variables.
Defines the possible values for an integer hyperparameter.
Defines the possible values for categorical, continuous, and integer hyperparameters to be used by an algorithm.
Defines a hyperparameter to be used by an algorithm.
Defines how the algorithm is used for a training job.
Create a machine learning algorithm that you can use in SageMaker and list in the Amazon Web Services Marketplace.
Create a machine learning algorithm that you can use in SageMaker and list in the Amazon Web Services Marketplace.
Creates a configuration for running a SageMaker AI image as a KernelGateway app. The configuration specifies the Amazon Elastic File System storage volume on the image, and a list of the kernels in the image.
Creates a configuration for running a SageMaker AI image as a KernelGateway app. The configuration specifies the Amazon Elastic File System storage volume on the image, and a list of the kernels in the image.
Creates a running app for the specified UserProfile. This operation is automatically invoked by Amazon SageMaker AI upon access to the associated Domain, and when new kernel configurations are selected by the user. A user may have multiple Apps active simultaneously.
Creates a running app for the specified UserProfile. This operation is automatically invoked by Amazon SageMaker AI upon access to the associated Domain, and when new kernel configurations are selected by the user. A user may have multiple Apps active simultaneously.
Creates an artifact. An artifact is a lineage tracking entity that represents a URI addressable object or data. Some examples are the S3 URI of a dataset and the ECR registry path of an image. For more information, see Amazon SageMaker ML Lineage Tracking.
Creates an artifact. An artifact is a lineage tracking entity that represents a URI addressable object or data. Some examples are the S3 URI of a dataset and the ECR registry path of an image. For more information, see Amazon SageMaker ML Lineage Tracking.
Specifies how to generate the endpoint name for an automatic one-click Autopilot model deployment.
Creates an Autopilot job also referred to as Autopilot experiment or AutoML job. An AutoML job in SageMaker AI is a fully automated process that allows you to build machine learning models with minimal effort and machine learning expertise. When initiating an AutoML job, you provide your data and optionally specify parameters tailored to your use case. SageMaker AI then automates the entire model development lifecycle, including data preprocessing, model training, tuning, and evaluation. AutoML jobs are designed to simplify and accelerate the model building process by automating various tasks and exploring different combinations of machine learning algorithms, data preprocessing techniques, and hyperparameter values. The output of an AutoML job comprises one or more trained models ready for deployment and inference. Additionally, SageMaker AI AutoML jobs generate a candidate model leaderboard, allowing you to select the best-performing model for deployment. For more information about AutoML jobs, see https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-automate-model-development.html in the SageMaker AI developer guide. We recommend using the new versions CreateAutoMLJobV2 and DescribeAutoMLJobV2, which offer backward compatibility. CreateAutoMLJobV2 can manage tabular problem types identical to those of its previous version CreateAutoMLJob, as well as time-series forecasting, non-tabular problem types such as image or text classification, and text generation (LLMs fine-tuning). Find guidelines about how to migrate a CreateAutoMLJob to CreateAutoMLJobV2 in Migrate a CreateAutoMLJob to CreateAutoMLJobV2. You can find the best-performing model after you run an AutoML job by calling DescribeAutoMLJobV2 (recommended) or DescribeAutoMLJob.
Creates an Autopilot job also referred to as Autopilot experiment or AutoML job. An AutoML job in SageMaker AI is a fully automated process that allows you to build machine learning models with minimal effort and machine learning expertise. When initiating an AutoML job, you provide your data and optionally specify parameters tailored to your use case. SageMaker AI then automates the entire model development lifecycle, including data preprocessing, model training, tuning, and evaluation. AutoML jobs are designed to simplify and accelerate the model building process by automating various tasks and exploring different combinations of machine learning algorithms, data preprocessing techniques, and hyperparameter values. The output of an AutoML job comprises one or more trained models ready for deployment and inference. Additionally, SageMaker AI AutoML jobs generate a candidate model leaderboard, allowing you to select the best-performing model for deployment. For more information about AutoML jobs, see https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-automate-model-development.html in the SageMaker AI developer guide. We recommend using the new versions CreateAutoMLJobV2 and DescribeAutoMLJobV2, which offer backward compatibility. CreateAutoMLJobV2 can manage tabular problem types identical to those of its previous version CreateAutoMLJob, as well as time-series forecasting, non-tabular problem types such as image or text classification, and text generation (LLMs fine-tuning). Find guidelines about how to migrate a CreateAutoMLJob to CreateAutoMLJobV2 in Migrate a CreateAutoMLJob to CreateAutoMLJobV2. You can find the best-performing model after you run an AutoML job by calling DescribeAutoMLJobV2 (recommended) or DescribeAutoMLJob.
Creates an Autopilot job also referred to as Autopilot experiment or AutoML job V2. An AutoML job in SageMaker AI is a fully automated process that allows you to build machine learning models with minimal effort and machine learning expertise. When initiating an AutoML job, you provide your data and optionally specify parameters tailored to your use case. SageMaker AI then automates the entire model development lifecycle, including data preprocessing, model training, tuning, and evaluation. AutoML jobs are designed to simplify and accelerate the model building process by automating various tasks and exploring different combinations of machine learning algorithms, data preprocessing techniques, and hyperparameter values. The output of an AutoML job comprises one or more trained models ready for deployment and inference. Additionally, SageMaker AI AutoML jobs generate a candidate model leaderboard, allowing you to select the best-performing model for deployment. For more information about AutoML jobs, see https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-automate-model-development.html in the SageMaker AI developer guide. AutoML jobs V2 support various problem types such as regression, binary, and multiclass classification with tabular data, text and image classification, time-series forecasting, and fine-tuning of large language models (LLMs) for text generation. CreateAutoMLJobV2 and DescribeAutoMLJobV2 are new versions of CreateAutoMLJob and DescribeAutoMLJob which offer backward compatibility. CreateAutoMLJobV2 can manage tabular problem types identical to those of its previous version CreateAutoMLJob, as well as time-series forecasting, non-tabular problem types such as image or text classification, and text generation (LLMs fine-tuning). Find guidelines about how to migrate a CreateAutoMLJob to CreateAutoMLJobV2 in Migrate a CreateAutoMLJob to CreateAutoMLJobV2. For the list of available problem types supported by CreateAutoMLJobV2, see AutoMLProblemTypeConfig. You can find the best-performing model after you run an AutoML job V2 by calling DescribeAutoMLJobV2.
Creates an Autopilot job also referred to as Autopilot experiment or AutoML job V2. An AutoML job in SageMaker AI is a fully automated process that allows you to build machine learning models with minimal effort and machine learning expertise. When initiating an AutoML job, you provide your data and optionally specify parameters tailored to your use case. SageMaker AI then automates the entire model development lifecycle, including data preprocessing, model training, tuning, and evaluation. AutoML jobs are designed to simplify and accelerate the model building process by automating various tasks and exploring different combinations of machine learning algorithms, data preprocessing techniques, and hyperparameter values. The output of an AutoML job comprises one or more trained models ready for deployment and inference. Additionally, SageMaker AI AutoML jobs generate a candidate model leaderboard, allowing you to select the best-performing model for deployment. For more information about AutoML jobs, see https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-automate-model-development.html in the SageMaker AI developer guide. AutoML jobs V2 support various problem types such as regression, binary, and multiclass classification with tabular data, text and image classification, time-series forecasting, and fine-tuning of large language models (LLMs) for text generation. CreateAutoMLJobV2 and DescribeAutoMLJobV2 are new versions of CreateAutoMLJob and DescribeAutoMLJob which offer backward compatibility. CreateAutoMLJobV2 can manage tabular problem types identical to those of its previous version CreateAutoMLJob, as well as time-series forecasting, non-tabular problem types such as image or text classification, and text generation (LLMs fine-tuning). Find guidelines about how to migrate a CreateAutoMLJob to CreateAutoMLJobV2 in Migrate a CreateAutoMLJob to CreateAutoMLJobV2. For the list of available problem types supported by CreateAutoMLJobV2, see AutoMLProblemTypeConfig. You can find the best-performing model after you run an AutoML job V2 by calling DescribeAutoMLJobV2.
Creates an Amazon SageMaker HyperPod cluster. SageMaker HyperPod is a capability of SageMaker for creating and managing persistent clusters for developing large machine learning models, such as large language models (LLMs) and diffusion models. To learn more, see Amazon SageMaker HyperPod in the Amazon SageMaker Developer Guide.
Creates an Amazon SageMaker HyperPod cluster. SageMaker HyperPod is a capability of SageMaker for creating and managing persistent clusters for developing large machine learning models, such as large language models (LLMs) and diffusion models. To learn more, see Amazon SageMaker HyperPod in the Amazon SageMaker Developer Guide.
Priority class configuration. When included in PriorityClasses, these class configurations define how tasks are queued.
Cluster policy configuration. This policy is used for task prioritization and fair-share allocation. This helps prioritize critical workloads and distributes idle compute across entities.
Create cluster policy configuration. This policy is used for task prioritization and fair-share allocation of idle compute. This helps prioritize critical workloads and distributes idle compute across entities.
Create cluster policy configuration. This policy is used for task prioritization and fair-share allocation of idle compute. This helps prioritize critical workloads and distributes idle compute across entities.
Creates a Git repository as a resource in your SageMaker AI account. You can associate the repository with notebook instances so that you can use Git source control for the notebooks you create. The Git repository is a resource in your SageMaker AI account, so it can be associated with more than one notebook instance, and it persists independently from the lifecycle of any notebook instances it is associated with. The repository can be hosted either in Amazon Web Services CodeCommit or in any other Git repository.
Creates a Git repository as a resource in your SageMaker AI account. You can associate the repository with notebook instances so that you can use Git source control for the notebooks you create. The Git repository is a resource in your SageMaker AI account, so it can be associated with more than one notebook instance, and it persists independently from the lifecycle of any notebook instances it is associated with. The repository can be hosted either in Amazon Web Services CodeCommit or in any other Git repository.
Contains information about a target platform that you want your model to run on, such as OS, architecture, and accelerators. It is an alternative of TargetDevice.
Contains information about the output location for the compiled model and the target device that the model runs on. TargetDevice and TargetPlatform are mutually exclusive, so you need to choose one between the two to specify your target device or platform. If you cannot find your device you want to use from the TargetDevice list, use TargetPlatform to describe the platform of your edge device and CompilerOptions if there are specific settings that are required or recommended to use for particular TargetPlatform.
The VpcConfig configuration object that specifies the VPC that you want the compilation jobs to connect to. For more information on controlling access to your Amazon S3 buckets used for compilation job, see Give Amazon SageMaker AI Compilation Jobs Access to Resources in Your Amazon VPC.
Contains information about the location of input model artifacts, the name and shape of the expected data inputs, and the framework in which the model was trained.
Starts a model compilation job. After the model has been compiled, Amazon SageMaker AI saves the resulting model artifacts to an Amazon Simple Storage Service (Amazon S3) bucket that you specify. If you choose to host your model using Amazon SageMaker AI hosting services, you can use the resulting model artifacts as part of the model. You can also use the artifacts with Amazon Web Services IoT Greengrass. In that case, deploy them as an ML resource. In the request body, you provide the following: A name for the compilation job Information about the input model artifacts The output location for the compiled model and the device (target) that the model runs on The Amazon Resource Name (ARN) of the IAM role that Amazon SageMaker AI assumes to perform the model compilation job. You can also provide a Tag to track the model compilation job's resource use and costs. The response body contains the CompilationJobArn for the compiled job. To stop a model compilation job, use StopCompilationJob. To get information about a particular model compilation job, use DescribeCompilationJob. To get information about multiple model compilation jobs, use ListCompilationJobs.
Starts a model compilation job. After the model has been compiled, Amazon SageMaker AI saves the resulting model artifacts to an Amazon Simple Storage Service (Amazon S3) bucket that you specify. If you choose to host your model using Amazon SageMaker AI hosting services, you can use the resulting model artifacts as part of the model. You can also use the artifacts with Amazon Web Services IoT Greengrass. In that case, deploy them as an ML resource. In the request body, you provide the following: A name for the compilation job Information about the input model artifacts The output location for the compiled model and the device (target) that the model runs on The Amazon Resource Name (ARN) of the IAM role that Amazon SageMaker AI assumes to perform the model compilation job. You can also provide a Tag to track the model compilation job's resource use and costs. The response body contains the CompilationJobArn for the compiled job. To stop a model compilation job, use StopCompilationJob. To get information about a particular model compilation job, use DescribeCompilationJob. To get information about multiple model compilation jobs, use ListCompilationJobs.
Create compute allocation definition. This defines how compute is allocated, shared, and borrowed for specified entities. Specifically, how to lend and borrow idle compute and assign a fair-share weight to the specified entities.
Create compute allocation definition. This defines how compute is allocated, shared, and borrowed for specified entities. Specifically, how to lend and borrow idle compute and assign a fair-share weight to the specified entities.
Creates a context. A context is a lineage tracking entity that represents a logical grouping of other tracking or experiment entities. Some examples are an endpoint and a model package. For more information, see Amazon SageMaker ML Lineage Tracking.
Creates a context. A context is a lineage tracking entity that represents a logical grouping of other tracking or experiment entities. Some examples are an endpoint and a model package. For more information, see Amazon SageMaker ML Lineage Tracking.
A time limit for how long the monitoring job is allowed to run before stopping.
Configuration for the cluster used to run model monitoring jobs.
Identifies the resources to deploy for a monitoring job.
Information about where and how you want to store the results of a monitoring job.
The output object for a monitoring job.
The output configuration for monitoring jobs.
The networking configuration for the monitoring job.
Input object for the endpoint
The input for the data quality monitoring job. Currently endpoints are supported for input.
The statistics resource for a monitoring job.
The constraints resource for a monitoring job.
Configuration for monitoring constraints and monitoring statistics. These baseline resources are compared against the results of the current job from the series of jobs scheduled to collect data periodically.
Information about the container that a data quality monitoring job runs.
Creates a definition for a job that monitors data quality and drift. For information about model monitor, see Amazon SageMaker AI Model Monitor.
Creates a definition for a job that monitors data quality and drift. For information about model monitor, see Amazon SageMaker AI Model Monitor.
The output configuration.
Creates a device fleet.
The TensorBoard app settings.
The SageMaker images that are hidden from the Studio user interface. You must specify the SageMaker image name and version aliases.
Studio settings. If these settings are applied on a user level, they take priority over the settings applied on a domain level.
Specifies options for sharing Amazon SageMaker AI Studio notebooks. These settings are specified as part of DefaultUserSettings when the CreateDomain API is called, and as part of UserSettings when the CreateUserProfile API is called. When SharingSettings is not specified, notebook sharing isn't allowed.
A collection of settings that configure user interaction with the RStudioServerPro app.
A collection of settings that apply to an RSessionGateway app.
The KernelGateway app settings.
The JupyterServer app settings.
The configuration parameters that specify the IAM roles assumed by the execution role of SageMaker (assumable roles) and the cluster instances or job execution environments (execution roles or runtime roles) to manage and access resources required for running Amazon EMR clusters or Amazon EMR Serverless applications.
The settings for the JupyterLab application.
A collection of default EBS storage settings that apply to spaces created within a domain or user profile.
The default storage settings for a space.
Details about the POSIX identity that is used for file system operations.
Configuration for the custom Amazon S3 file system.
The settings for assigning a custom Amazon FSx for Lustre file system to a user profile or space for an Amazon SageMaker Domain.
The settings for assigning a custom Amazon EFS file system to a user profile or space for an Amazon SageMaker AI Domain.
The settings for assigning a custom file system to a user profile or space for an Amazon SageMaker AI Domain. Permitted users can access this file system in Amazon SageMaker AI Studio.
A collection of settings that apply to users in a domain. These settings are specified when the CreateUserProfile API is called, and as DefaultUserSettings when the CreateDomain API is called. SecurityGroups is aggregated when specified in both calls. For all other settings in UserSettings, the values specified in CreateUserProfile take precedence over those specified in CreateDomain.
The settings that apply to an Amazon SageMaker AI domain when you use it in Amazon SageMaker Unified Studio.
The Trusted Identity Propagation (TIP) settings for the SageMaker domain. These settings determine how user identities from IAM Identity Center are propagated through the domain to TIP enabled Amazon Web Services services.
A collection of settings that configure the RStudioServerPro Domain-level app.
A collection of settings that configure the domain's Docker interaction.
A collection of settings that apply to the SageMaker Domain. These settings are specified through the CreateDomain API call.
The default settings for shared spaces that users create in the domain. SageMaker applies these settings only to shared spaces. It doesn't apply them to private spaces.
Creates a Domain. A domain consists of an associated Amazon Elastic File System volume, a list of authorized users, and a variety of security, application, policy, and Amazon Virtual Private Cloud (VPC) configurations. Users within a domain can share notebook files and other artifacts with each other. EFS storage When a domain is created, an EFS volume is created for use by all of the users within the domain. Each user receives a private home directory within the EFS volume for notebooks, Git repositories, and data files. SageMaker AI uses the Amazon Web Services Key Management Service (Amazon Web Services KMS) to encrypt the EFS volume attached to the domain with an Amazon Web Services managed key by default. For more control, you can specify a customer managed key. For more information, see Protect Data at Rest Using Encryption. VPC configuration All traffic between the domain and the Amazon EFS volume is through the specified VPC and subnets. For other traffic, you can specify the AppNetworkAccessType parameter. AppNetworkAccessType corresponds to the network access type that you choose when you onboard to the domain. The following options are available: PublicInternetOnly - Non-EFS traffic goes through a VPC managed by Amazon SageMaker AI, which allows internet access. This is the default value. VpcOnly - All traffic is through the specified VPC and subnets. Internet access is disabled by default. To allow internet access, you must specify a NAT gateway. When internet access is disabled, you won't be able to run a Amazon SageMaker AI Studio notebook or to train or host models unless your VPC has an interface endpoint to the SageMaker AI API and runtime or a NAT gateway and your security groups allow outbound connections. NFS traffic over TCP on port 2049 needs to be allowed in both inbound and outbound rules in order to launch a Amazon SageMaker AI Studio app successfully. For more information, see Connect Amazon SageMaker AI Studio Notebooks to Resources in a VPC.
Creates a Domain. A domain consists of an associated Amazon Elastic File System volume, a list of authorized users, and a variety of security, application, policy, and Amazon Virtual Private Cloud (VPC) configurations. Users within a domain can share notebook files and other artifacts with each other. EFS storage When a domain is created, an EFS volume is created for use by all of the users within the domain. Each user receives a private home directory within the EFS volume for notebooks, Git repositories, and data files. SageMaker AI uses the Amazon Web Services Key Management Service (Amazon Web Services KMS) to encrypt the EFS volume attached to the domain with an Amazon Web Services managed key by default. For more control, you can specify a customer managed key. For more information, see Protect Data at Rest Using Encryption. VPC configuration All traffic between the domain and the Amazon EFS volume is through the specified VPC and subnets. For other traffic, you can specify the AppNetworkAccessType parameter. AppNetworkAccessType corresponds to the network access type that you choose when you onboard to the domain. The following options are available: PublicInternetOnly - Non-EFS traffic goes through a VPC managed by Amazon SageMaker AI, which allows internet access. This is the default value. VpcOnly - All traffic is through the specified VPC and subnets. Internet access is disabled by default. To allow internet access, you must specify a NAT gateway. When internet access is disabled, you won't be able to run a Amazon SageMaker AI Studio notebook or to train or host models unless your VPC has an interface endpoint to the SageMaker AI API and runtime or a NAT gateway and your security groups allow outbound connections. NFS traffic over TCP on port 2049 needs to be allowed in both inbound and outbound rules in order to launch a Amazon SageMaker AI Studio app successfully. For more information, see Connect Amazon SageMaker AI Studio Notebooks to Resources in a VPC.
Contains information about the configuration of a model in a deployment.
Contains information about the configuration of a deployment.
Contains information about the configurations of selected devices.
Contains information about a stage in an edge deployment plan.
Creates an edge deployment plan, consisting of multiple stages. Each stage may have a different deployment configuration and devices.
Creates an edge deployment plan, consisting of multiple stages. Each stage may have a different deployment configuration and devices.
Creates a new stage in an existing edge deployment plan.
Starts a SageMaker Edge Manager model packaging job. Edge Manager will use the model artifacts from the Amazon Simple Storage Service bucket that you specify. After the model has been packaged, Amazon SageMaker saves the resulting artifacts to an S3 bucket that you specify.
Specifies the serverless configuration for an endpoint variant.
Settings that control how the endpoint routes incoming traffic to the instances that the endpoint hosts.
Configures the scale-in behavior for managed instance scaling.
Settings that control the range in the number of instances that the endpoint provisions as it scales up or down to accommodate traffic.
Specifies configuration for a core dump from the model container when the process crashes.
Settings for the capacity reservation for the compute instances that SageMaker AI reserves for an endpoint.
Specifies an instance type and its priority for a heterogeneous endpoint. Use instance pools to configure a production variant with multiple instance types, enabling the endpoint to provision instances across different types based on priority.
Identifies a model that you want to host and the resources chosen to deploy for hosting it. If you are deploying multiple models, tell SageMaker how to distribute traffic among the models by specifying variant weights. For more information on production variants, check Production variants.
The configuration for Utilization metrics.
A parameter to activate explainers.
Configuration to control how SageMaker AI captures inference data.
Creates an endpoint configuration that SageMaker hosting services uses to deploy models. In the configuration, you identify one or more models, created using the CreateModel API, to deploy and the resources that you want SageMaker to provision. Then you call the CreateEndpoint API. Use this API if you want to use SageMaker hosting services to deploy models into production. In the request, you define a ProductionVariant, for each model that you want to deploy. Each ProductionVariant parameter also describes the resources that you want SageMaker to provision. This includes the number and type of ML compute instances to deploy. If you are hosting multiple models, you also assign a VariantWeight to specify how much traffic you want to allocate to each model. For example, suppose that you want to host two models, A and B, and you assign traffic weight 2 for model A and 1 for model B. SageMaker distributes two-thirds of the traffic to Model A, and one-third to model B. When you call CreateEndpoint, a load call is made to DynamoDB to verify that your endpoint configuration exists. When you read data from a DynamoDB table supporting Eventually Consistent Reads , the response might not reflect the results of a recently completed write operation. The response might include some stale data. If the dependent entities are not yet in DynamoDB, this causes a validation error. If you repeat your read request after a short time, the response should return the latest data. So retry logic is recommended to handle these possible issues. We also recommend that customers call DescribeEndpointConfig before calling CreateEndpoint to minimize the potential impact of a DynamoDB eventually consistent read.
Creates an endpoint configuration that SageMaker hosting services uses to deploy models. In the configuration, you identify one or more models, created using the CreateModel API, to deploy and the resources that you want SageMaker to provision. Then you call the CreateEndpoint API. Use this API if you want to use SageMaker hosting services to deploy models into production. In the request, you define a ProductionVariant, for each model that you want to deploy. Each ProductionVariant parameter also describes the resources that you want SageMaker to provision. This includes the number and type of ML compute instances to deploy. If you are hosting multiple models, you also assign a VariantWeight to specify how much traffic you want to allocate to each model. For example, suppose that you want to host two models, A and B, and you assign traffic weight 2 for model A and 1 for model B. SageMaker distributes two-thirds of the traffic to Model A, and one-third to model B. When you call CreateEndpoint, a load call is made to DynamoDB to verify that your endpoint configuration exists. When you read data from a DynamoDB table supporting Eventually Consistent Reads , the response might not reflect the results of a recently completed write operation. The response might include some stale data. If the dependent entities are not yet in DynamoDB, this causes a validation error. If you repeat your read request after a short time, the response should return the latest data. So retry logic is recommended to handle these possible issues. We also recommend that customers call DescribeEndpointConfig before calling CreateEndpoint to minimize the potential impact of a DynamoDB eventually consistent read.
Specifies a rolling deployment strategy for updating a SageMaker endpoint.
The deployment configuration for an endpoint, which contains the desired deployment strategy and rollback configurations.
Creates an endpoint using the endpoint configuration specified in the request. SageMaker uses the endpoint to provision resources and deploy models. You create the endpoint configuration with the CreateEndpointConfig API. Use this API to deploy models using SageMaker hosting services. You must not delete an EndpointConfig that is in use by an endpoint that is live or while the UpdateEndpoint or CreateEndpoint operations are being performed on the endpoint. To update an endpoint, you must create a new EndpointConfig. The endpoint name must be unique within an Amazon Web Services Region in your Amazon Web Services account. When it receives the request, SageMaker creates the endpoint, launches the resources (ML compute instances), and deploys the model(s) on them. When you call CreateEndpoint, a load call is made to DynamoDB to verify that your endpoint configuration exists. When you read data from a DynamoDB table supporting Eventually Consistent Reads , the response might not reflect the results of a recently completed write operation. The response might include some stale data. If the dependent entities are not yet in DynamoDB, this causes a validation error. If you repeat your read request after a short time, the response should return the latest data. So retry logic is recommended to handle these possible issues. We also recommend that customers call DescribeEndpointConfig before calling CreateEndpoint to minimize the potential impact of a DynamoDB eventually consistent read. When SageMaker receives the request, it sets the endpoint status to Creating. After it creates the endpoint, it sets the status to InService. SageMaker can then process incoming requests for inferences. To check the status of an endpoint, use the DescribeEndpoint API. If any of the models hosted at this endpoint get model data from an Amazon S3 location, SageMaker uses Amazon Web Services Security Token Service to download model artifacts from the S3 path you provided. Amazon Web Services STS is activated in your Amazon Web Services account by default. If you previously deactivated Amazon Web Services STS for a region, you need to reactivate Amazon Web Services STS for that region. For more information, see Activating and Deactivating Amazon Web Services STS in an Amazon Web Services Region in the Amazon Web Services Identity and Access Management User Guide. To add the IAM role policies for using this API operation, go to the IAM console, and choose Roles in the left navigation pane. Search the IAM role that you want to grant access to use the CreateEndpoint and CreateEndpointConfig API operations, add the following policies to the role. Option 1: For a full SageMaker access, search and attach the AmazonSageMakerFullAccess policy. Option 2: For granting a limited access to an IAM role, paste the following Action elements manually into the JSON file of the IAM role: "Action": ["sagemaker:CreateEndpoint", "sagemaker:CreateEndpointConfig"] "Resource": [ "arn:aws:sagemaker:region:account-id:endpoint/endpointName" "arn:aws:sagemaker:region:account-id:endpoint-config/endpointConfigName" ] For more information, see SageMaker API Permissions: Actions, Permissions, and Resources Reference.
Creates an endpoint using the endpoint configuration specified in the request. SageMaker uses the endpoint to provision resources and deploy models. You create the endpoint configuration with the CreateEndpointConfig API. Use this API to deploy models using SageMaker hosting services. You must not delete an EndpointConfig that is in use by an endpoint that is live or while the UpdateEndpoint or CreateEndpoint operations are being performed on the endpoint. To update an endpoint, you must create a new EndpointConfig. The endpoint name must be unique within an Amazon Web Services Region in your Amazon Web Services account. When it receives the request, SageMaker creates the endpoint, launches the resources (ML compute instances), and deploys the model(s) on them. When you call CreateEndpoint, a load call is made to DynamoDB to verify that your endpoint configuration exists. When you read data from a DynamoDB table supporting Eventually Consistent Reads , the response might not reflect the results of a recently completed write operation. The response might include some stale data. If the dependent entities are not yet in DynamoDB, this causes a validation error. If you repeat your read request after a short time, the response should return the latest data. So retry logic is recommended to handle these possible issues. We also recommend that customers call DescribeEndpointConfig before calling CreateEndpoint to minimize the potential impact of a DynamoDB eventually consistent read. When SageMaker receives the request, it sets the endpoint status to Creating. After it creates the endpoint, it sets the status to InService. SageMaker can then process incoming requests for inferences. To check the status of an endpoint, use the DescribeEndpoint API. If any of the models hosted at this endpoint get model data from an Amazon S3 location, SageMaker uses Amazon Web Services Security Token Service to download model artifacts from the S3 path you provided. Amazon Web Services STS is activated in your Amazon Web Services account by default. If you previously deactivated Amazon Web Services STS for a region, you need to reactivate Amazon Web Services STS for that region. For more information, see Activating and Deactivating Amazon Web Services STS in an Amazon Web Services Region in the Amazon Web Services Identity and Access Management User Guide. To add the IAM role policies for using this API operation, go to the IAM console, and choose Roles in the left navigation pane. Search the IAM role that you want to grant access to use the CreateEndpoint and CreateEndpointConfig API operations, add the following policies to the role. Option 1: For a full SageMaker access, search and attach the AmazonSageMakerFullAccess policy. Option 2: For granting a limited access to an IAM role, paste the following Action elements manually into the JSON file of the IAM role: "Action": ["sagemaker:CreateEndpoint", "sagemaker:CreateEndpointConfig"] "Resource": [ "arn:aws:sagemaker:region:account-id:endpoint/endpointName" "arn:aws:sagemaker:region:account-id:endpoint-config/endpointConfigName" ] For more information, see SageMaker API Permissions: Actions, Permissions, and Resources Reference.
Creates a SageMaker experiment. An experiment is a collection of trials that are observed, compared and evaluated as a group. A trial is a set of steps, called trial components, that produce a machine learning model. In the Studio UI, trials are referred to as run groups and trial components are referred to as runs. The goal of an experiment is to determine the components that produce the best model. Multiple trials are performed, each one isolating and measuring the impact of a change to one or more inputs, while keeping the remaining inputs constant. When you use SageMaker Studio or the SageMaker Python SDK, all experiments, trials, and trial components are automatically tracked, logged, and indexed. When you use the Amazon Web Services SDK for Python (Boto), you must use the logging APIs provided by the SDK. You can add tags to experiments, trials, trial components and then use the Search API to search for the tags. To add a description to an experiment, specify the optional Description parameter. To add a description later, or to change the description, call the UpdateExperiment API. To get a list of all your experiments, call the ListExperiments API. To view an experiment's properties, call the DescribeExperiment API. To get a list of all the trials associated with an experiment, call the ListTrials API. To create a trial call the CreateTrial API.
Creates a SageMaker experiment. An experiment is a collection of trials that are observed, compared and evaluated as a group. A trial is a set of steps, called trial components, that produce a machine learning model. In the Studio UI, trials are referred to as run groups and trial components are referred to as runs. The goal of an experiment is to determine the components that produce the best model. Multiple trials are performed, each one isolating and measuring the impact of a change to one or more inputs, while keeping the remaining inputs constant. When you use SageMaker Studio or the SageMaker Python SDK, all experiments, trials, and trial components are automatically tracked, logged, and indexed. When you use the Amazon Web Services SDK for Python (Boto), you must use the logging APIs provided by the SDK. You can add tags to experiments, trials, trial components and then use the Search API to search for the tags. To add a description to an experiment, specify the optional Description parameter. To add a description later, or to change the description, call the UpdateExperiment API. To get a list of all your experiments, call the ListExperiments API. To view an experiment's properties, call the DescribeExperiment API. To get a list of all the trials associated with an experiment, call the ListTrials API. To create a trial call the CreateTrial API.
Used to set feature group throughput configuration. There are two modes: ON_DEMAND and PROVISIONED. With on-demand mode, you are charged for data reads and writes that your application performs on your feature group. You do not need to specify read and write throughput because Feature Store accommodates your workloads as they ramp up and down. You can switch a feature group to on-demand only once in a 24 hour period. With provisioned throughput mode, you specify the read and write capacity per second that you expect your application to require, and you are billed based on those limits. Exceeding provisioned throughput will result in your requests being throttled. Note: PROVISIONED throughput mode is supported only for feature groups that are offline-only, or use the Standard tier online store.
Time to live duration, where the record is hard deleted after the expiration time is reached; ExpiresAt = EventTime + TtlDuration. For information on HardDelete, see the DeleteRecord API in the Amazon SageMaker API Reference guide.
The security configuration for OnlineStore.
Use this to specify the Amazon Web Services Key Management Service (KMS) Key ID, or KMSKeyId, for at rest data encryption. You can turn OnlineStore on or off by specifying the EnableOnlineStore flag at General Assembly. The default value is False.
The Amazon Simple Storage (Amazon S3) location and security configuration for OfflineStore.
The meta data of the Glue table which serves as data catalog for the OfflineStore.
The configuration of an OfflineStore. Provide an OfflineStoreConfig in a request to CreateFeatureGroup to create an OfflineStore. To encrypt an OfflineStore using at rest data encryption, specify Amazon Web Services Key Management Service (KMS) key ID, or KMSKeyId, in S3StorageConfig.
A list of features. You must include FeatureName and FeatureType. Valid feature FeatureTypes are Integral, Fractional and String.
Create a new FeatureGroup. A FeatureGroup is a group of Features defined in the FeatureStore to describe a Record. The FeatureGroup defines the schema and features contained in the FeatureGroup. A FeatureGroup definition is composed of a list of Features, a RecordIdentifierFeatureName, an EventTimeFeatureName and configurations for its OnlineStore and OfflineStore. Check Amazon Web Services service quotas to see the FeatureGroups quota for your Amazon Web Services account. Note that it can take approximately 10-15 minutes to provision an OnlineStore FeatureGroup with the InMemory StorageType. You must include at least one of OnlineStoreConfig and OfflineStoreConfig to create a FeatureGroup.
Create a new FeatureGroup. A FeatureGroup is a group of Features defined in the FeatureStore to describe a Record. The FeatureGroup defines the schema and features contained in the FeatureGroup. A FeatureGroup definition is composed of a list of Features, a RecordIdentifierFeatureName, an EventTimeFeatureName and configurations for its OnlineStore and OfflineStore. Check Amazon Web Services service quotas to see the FeatureGroups quota for your Amazon Web Services account. Note that it can take approximately 10-15 minutes to provision an OnlineStore FeatureGroup with the InMemory StorageType. You must include at least one of OnlineStoreConfig and OfflineStoreConfig to create a FeatureGroup.
Container for configuring the source of human task requests.
Defines the amount of money paid to an Amazon Mechanical Turk worker for each task performed. Use one of the following prices for bounding box tasks. Prices are in US dollars and should be based on the complexity of the task; the longer it takes in your initial testing, the more you should offer. 0.036 0.048 0.060 0.072 0.120 0.240 0.360 0.480 0.600 0.720 0.840 0.960 1.080 1.200 Use one of the following prices for image classification, text classification, and custom tasks. Prices are in US dollars. 0.012 0.024 0.036 0.048 0.060 0.072 0.120 0.240 0.360 0.480 0.600 0.720 0.840 0.960 1.080 1.200 Use one of the following prices for semantic segmentation tasks. Prices are in US dollars. 0.840 0.960 1.080 1.200 Use one of the following prices for Textract AnalyzeDocument Important Form Key Amazon Augmented AI review tasks. Prices are in US dollars. 2.400 2.280 2.160 2.040 1.920 1.800 1.680 1.560 1.440 1.320 1.200 1.080 0.960 0.840 0.720 0.600 0.480 0.360 0.240 0.120 0.072 0.060 0.048 0.036 0.024 0.012 Use one of the following prices for Rekognition DetectModerationLabels Amazon Augmented AI review tasks. Prices are in US dollars. 1.200 1.080 0.960 0.840 0.720 0.600 0.480 0.360 0.240 0.120 0.072 0.060 0.048 0.036 0.024 0.012 Use one of the following prices for Amazon Augmented AI custom human review tasks. Prices are in US dollars. 1.200 1.080 0.960 0.840 0.720 0.600 0.480 0.360 0.240 0.120 0.072 0.060 0.048 0.036 0.024 0.012
Describes the work to be performed by human workers.
Defines under what conditions SageMaker creates a human loop. Used within CreateFlowDefinition. See HumanLoopActivationConditionsConfig for the required format of activation conditions.
Provides information about how and under what conditions SageMaker creates a human loop. If HumanLoopActivationConfig is not given, then all requests go to humans.
Contains information about where human output will be stored.
Creates a flow definition.
Creates a flow definition.
Configuration for accessing hub content through presigned URLs, including license agreement acceptance and URL validation settings.
Creates presigned URLs for accessing hub content artifacts. This operation generates time-limited, secure URLs that allow direct download of model artifacts and associated files from Amazon SageMaker hub content, including gated models that require end-user license agreement acceptance.
Creates presigned URLs for accessing hub content artifacts. This operation generates time-limited, secure URLs that allow direct download of model artifacts and associated files from Amazon SageMaker hub content, including gated models that require end-user license agreement acceptance.
Create a hub content reference in order to add a model in the JumpStart public hub to a private hub.
Create a hub content reference in order to add a model in the JumpStart public hub to a private hub.
The Amazon S3 storage configuration of a hub.
Create a hub.
Create a hub.
The Liquid template for the worker user interface.
Defines the settings you will use for the human review workflow user interface. Reviewers will see a three-panel interface with an instruction area, the item to review, and an input area.
Defines the settings you will use for the human review workflow user interface. Reviewers will see a three-panel interface with an instruction area, the item to review, and an input area.
A previously completed or stopped hyperparameter tuning job to be used as a starting point for a new hyperparameter tuning job.
Specifies the configuration for a hyperparameter tuning job that uses one or more previous hyperparameter tuning jobs as a starting point. The results of previous tuning jobs are used to inform which combinations of hyperparameters to search over in the new tuning job. All training jobs launched by the new hyperparameter tuning job are evaluated by using the objective metric, and the training job that performs the best is compared to the best training jobs from the parent tuning jobs. From these, the training job that performs the best as measured by the objective metric is returned as the overall best training job. All training jobs launched by parent hyperparameter tuning jobs and the new hyperparameter tuning jobs count against the limit of training jobs for the tuning job.
The job completion criteria.
Specifies the maximum number of training jobs and parallel training jobs that a hyperparameter tuning job can launch.
For a hyperparameter of the integer type, specifies the range that a hyperparameter tuning job searches.
Specifies ranges of integer, continuous, and categorical hyperparameters that a hyperparameter tuning job searches. The hyperparameter tuning job launches training jobs with hyperparameter values within these ranges to find the combination of values that result in the training job with the best performance as measured by the objective metric of the hyperparameter tuning job. The maximum number of items specified for Array Members refers to the maximum number of hyperparameters for each range and also the maximum for the hyperparameter tuning job itself. That is, the sum of the number of hyperparameters for all the ranges can't exceed the maximum number specified.
The configuration for Hyperband, a multi-fidelity based hyperparameter tuning strategy. Hyperband uses the final and intermediate results of a training job to dynamically allocate resources to utilized hyperparameter configurations while automatically stopping under-performing configurations. This parameter should be provided only if Hyperband is selected as the StrategyConfig under the HyperParameterTuningJobConfig API.
The configuration for a training job launched by a hyperparameter tuning job. Choose Bayesian for Bayesian optimization, and Random for random search optimization. For more advanced use cases, use Hyperband, which evaluates objective metrics for training jobs after every epoch. For more information about strategies, see How Hyperparameter Tuning Works.
Configures a hyperparameter tuning job.
The retry strategy to use when a training job fails due to an InternalServerError. RetryStrategy is specified as part of the CreateTrainingJob and CreateHyperParameterTuningJob requests. You can add the StoppingCondition parameter to the request to limit the training time for the complete job.
The configuration for hyperparameter tuning resources for use in training jobs launched by the tuning job. These resources include compute instances and storage volumes. Specify one or more compute instance configurations and allocation strategies to select resources (optional).
The configuration of resources, including compute instances and storage volumes for use in training jobs launched by hyperparameter tuning jobs. HyperParameterTuningResourceConfig is similar to ResourceConfig, but has the additional InstanceConfigs and AllocationStrategy fields to allow for flexible instance management. Specify one or more instance types, count, and the allocation strategy for instance selection. HyperParameterTuningResourceConfig supports the capabilities of ResourceConfig with the exception of KeepAlivePeriodInSeconds. Hyperparameter tuning jobs use warm pools by default, which reuse clusters between training jobs.