ModelData

class aws_cdk.aws_sagemaker_alpha.ModelData

Bases: object

(experimental) Model data represents the source of model artifacts, which will ultimately be loaded from an S3 location.

Stability:: experimental
ExampleMetadata:: infused

Example:

import aws_cdk.aws_sagemaker_alpha as sagemaker
import path as path


image = sagemaker.ContainerImage.from_asset(path.join("path", "to", "Dockerfile", "directory"))
model_data = sagemaker.ModelData.from_asset(path.join("path", "to", "artifact", "file.tar.gz"))

model = sagemaker.Model(self, "PrimaryContainerModel",
    containers=[sagemaker.ContainerDefinition(
        image=image,
        model_data=model_data
    )
    ]
)

Stability:: experimental

Methods

abstract bind(scope, model)

(experimental) This method is invoked by the SageMaker Model construct when it needs to resolve the model data to a URI.

Parameters:

scope (Construct) – The scope within which the model data is resolved.
model (IModel) – The Model construct performing the URI resolution.

Stability:

experimental

Return type:

ModelDataConfig

Static Methods

classmethod from_asset(path, *, deploy_time=None, readers=None, asset_hash=None, asset_hash_type=None, bundling=None, exclude=None, follow_symlinks=None, ignore_mode=None)

(experimental) Constructs model data that will be uploaded to S3 as part of the CDK app deployment.

Parameters:

path (str) – The local path to a model artifact file as a gzipped tar file.
deploy_time (Optional[bool]) – Whether or not the asset needs to exist beyond deployment time; i.e. are copied over to a different location and not needed afterwards. Setting this property to true has an impact on the lifecycle of the asset, because we will assume that it is safe to delete after the CloudFormation deployment succeeds. For example, Lambda Function assets are copied over to Lambda during deployment. Therefore, it is not necessary to store the asset in S3, so we consider those deployTime assets. Default: false
readers (Optional[Sequence[IGrantable]]) – A list of principals that should be able to read this asset from S3. You can use asset.grantRead(principal) to grant read permissions later. Default: - No principals that can read file asset.
asset_hash (Optional[str]) – Specify a custom hash for this asset. If assetHashType is set it must be set to AssetHashType.CUSTOM. For consistency, this custom hash will be SHA256 hashed and encoded as hex. The resulting hash will be the asset hash. NOTE: the hash is used in order to identify a specific revision of the asset, and used for optimizing and caching deployment activities related to this asset such as packaging, uploading to Amazon S3, etc. If you chose to customize the hash, you will need to make sure it is updated every time the asset changes, or otherwise it is possible that some deployments will not be invalidated. Default: - based on assetHashType
asset_hash_type (Optional[AssetHashType]) – Specifies the type of hash to calculate for this asset. If assetHash is configured, this option must be undefined or AssetHashType.CUSTOM. Default: - the default is AssetHashType.SOURCE, but if assetHash is explicitly specified this value defaults to AssetHashType.CUSTOM.
bundling (Union[BundlingOptions, Dict[str, Any], None]) – Bundle the asset by executing a command in a Docker container or a custom bundling provider. The asset path will be mounted at /asset-input. The Docker container is responsible for putting content at /asset-output. The content at /asset-output will be zipped and used as the final asset. Default: - uploaded as-is to S3 if the asset is a regular file or a .zip file, archived into a .zip file and uploaded to S3 otherwise
exclude (Optional[Sequence[str]]) – File paths matching the patterns will be excluded. See ignoreMode to set the matching behavior. Has no effect on Assets bundled using the bundling property. Default: - nothing is excluded
follow_symlinks (Optional[SymlinkFollowMode]) – A strategy for how to handle symlinks. Default: SymlinkFollowMode.NEVER
ignore_mode (Optional[IgnoreMode]) – The ignore behavior to use for exclude patterns. Default: IgnoreMode.GLOB

Stability:

experimental

Return type:

ModelData

classmethod from_bucket(bucket, object_key)

(experimental) Constructs model data which is already available within S3.

Parameters:

bucket (IBucket) – The S3 bucket within which the model artifacts are stored.
object_key (str) – The S3 object key at which the model artifacts are stored.

Stability:

experimental

Return type:

ModelData