Amazon Bedrock

{'customModelDataSource': {'modelPackageArnDataSource': {'modelPackageArn': 'string'}}}

Creates a new custom model in Amazon Bedrock. After the model is active, you can use it for inference.

You can provide the model data source in one of the following ways:

• customModelDataSource — Specify a SageMaker AI model package ARN. Amazon Bedrock resolves the model package to retrieve the model artifacts. This is the preferred method for new SageMaker AI training outputs.

• modelSourceConfig — Specify an Amazon S3 URI pointing to the Amazon-managed Amazon S3 bucket containing your model artifacts.

To use the model for inference, you must purchase Provisioned Throughput for it. You can't use On-demand inference with these custom models. For more information about Provisioned Throughput, see Provisioned Throughput.

The model appears in ListCustomModels with a customizationType of imported. To track the status of the new model, you use the GetCustomModel API operation. The model can be in the following states:

• Creating - Initial state during validation and registration

• Active - Model is ready for use in inference

• Failed - Creation process encountered an error

Related APIs

• GetCustomModel

• ListCustomModels

• DeleteCustomModel

See also: AWS API Documentation

Request Syntax

client.create_custom_model(
    modelName='string',
    modelSourceConfig={
        's3DataSource': {
            's3Uri': 'string'
        }
    },
    customModelDataSource={
        'modelPackageArnDataSource': {
            'modelPackageArn': 'string'
        }
    },
    modelKmsKeyArn='string',
    roleArn='string',
    modelTags=[
        {
            'key': 'string',
            'value': 'string'
        },
    ],
    clientRequestToken='string'
)

type modelName:

string

param modelName:

[REQUIRED]

A unique name for the custom model.

type modelSourceConfig:

dict

param modelSourceConfig:

The data source for the model. The Amazon S3 URI in the model source must be for the Amazon-managed Amazon S3 bucket containing your model artifacts.

• s3DataSource (dict) --

  The Amazon S3 data source of the model to import.

  • s3Uri (string) -- [REQUIRED]

    The URI of the Amazon S3 data source.

type customModelDataSource:

dict

param customModelDataSource:

The data source for the custom model. Use this field to specify a SageMaker AI model package ARN as the source for your custom model. Amazon Bedrock resolves the model package to retrieve the model artifacts.

You can specify either customModelDataSource or modelSourceConfig, but not both.

• modelPackageArnDataSource (dict) --

  A SageMaker AI model package ARN as the data source for the custom model. When you specify a model package ARN, Amazon Bedrock resolves the model package to retrieve the model artifacts.

  • modelPackageArn (string) -- [REQUIRED]

    The Amazon Resource Name (ARN) of the SageMaker AI model package. The ARN must be for a model package of restricted type.

    To use a model package ARN, you must have the sagemaker:DescribeModelPackage and sagemaker:AccessModelPackageData permissions on the model package resource.

type modelKmsKeyArn:

string

param modelKmsKeyArn:

The Amazon Resource Name (ARN) of the customer managed KMS key to encrypt the custom model. If you don't provide a KMS key, Amazon Bedrock uses an Amazon Web Services-managed KMS key to encrypt the model.

If you provide a customer managed KMS key, your Amazon Bedrock service role must have permissions to use it. For more information see Encryption of imported models.

type roleArn:

string

param roleArn:

The Amazon Resource Name (ARN) of an IAM service role that Amazon Bedrock assumes to perform tasks on your behalf. This role must have permissions to access the Amazon S3 bucket containing your model artifacts and the KMS key (if specified). For more information, see Setting up an IAM service role for importing models in the Amazon Bedrock User Guide.

This field is required when you use modelSourceConfig with an Amazon S3 data source. It is not required when you use customModelDataSource with a model package ARN, because Amazon Bedrock uses its own credentials to access the model artifacts.

type modelTags:

list

param modelTags:

A list of key-value pairs to associate with the custom model resource. You can use these tags to organize and identify your resources.

For more information, see Tagging resources in the Amazon Bedrock User Guide.

• (dict) --

  Definition of the key/value pair for a tag.

  • key (string) -- [REQUIRED]

    Key for the tag.

  • value (string) -- [REQUIRED]

    Value for the tag.

type clientRequestToken:

string

param clientRequestToken:

A unique, case-sensitive identifier to ensure that the API request completes no more than one time. If this token matches a previous request, Amazon Bedrock ignores the request, but does not return an error. For more information, see Ensuring idempotency.

This field is autopopulated if not provided.

rtype:

dict

returns:

Response Syntax

{
    'modelArn': 'string'
}

Response Structure

• (dict) --

  • modelArn (string) --

    The Amazon Resource Name (ARN) of the new custom model.