class google.cloud.aiplatform.metadata.schema.google.artifact_schema.ExperimentModel(*, framework_name:str, framework_version:str, model_file:str, uri:str, model_class:Optional[str] = None, predict_schemata:Optional[google.cloud.aiplatform.metadata.schema.utils.PredictSchemata] = None, artifact_id:Optional[str] = None, display_name:Optional[str] = None, schema_version:Optional[str] = None, description:Optional[str] = None, metadata:Optional[Dict] = None, state:Optional[google.cloud.aiplatform_v1.types.artifact.Artifact.State] = State.LIVE)

An artifact representing a Vertex Experiment Model.

Instantiates an ExperimentModel that represents a saved ML model.

• Parameters

  • framework_name (str) – Required. The name of the model’s framework. E.g., ‘sklearn’

  • framework_version (str) – Required. The version of the model’s framework. E.g., ‘1.1.0’

  • model_file (str) – Required. The file name of the model. E.g., ‘model.pkl’

  • uri (str) – Required. The uniform resource identifier of the model artifact directory.

  • model_class (str) – Optional. The class name of the model. E.g., ‘sklearn.linear_model._base.LinearRegression’

  • predict_schemata (PredictSchemata) – Optional. An instance of PredictSchemata which holds instance, parameter and prediction schema uris.

  • artifact_id (str) – Optional. The <resource_id> portion of the Artifact name withthe format. This is globally unique in a metadataStore:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/artifacts/<resource_id>.

  • display_name (str) – Optional. The user-defined name of the Artifact.

  • schema_version (str) – Optional. schema_version specifies the version used by the Artifact.If not set, defaults to use the latest version.

  • description (str) – Optional. Describes the purpose of the Artifact to be created.

  • metadata (Dict) – Optional. Contains the metadata information that will be stored in the Artifact.

  • state (google.cloud.gapic.types.Artifact.State) – Optional. The state of this Artifact. This is aproperty of the Artifact, and does not imply orapture any ongoing process. This property ismanaged by clients (such as Vertex AIPipelines), and the system does not prescribe orcheck the validity of state transitions.

property framework_name(: Optional[str )

The framework name of the saved ML model.

property framework_version(: Optional[str )

The framework version of the saved ML model.

classmethod get(artifact_id:str, *, metadata_store_id:str = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Retrieves an existing ExperimentModel artifact given an artifact id.

• Parameters

  • artifact_id (str) – Required. An artifact id of the ExperimentModel artifact.

  • metadata_store_id (str) – Optional. MetadataStore to retrieve Artifact from. If not set, metadata_store_id is set to “default”.If artifact_id is a fully-qualified resource name, its metadata_store_id overrides this one.

  • project (str) – Optional. Project to retrieve the artifact from. If not set, projectset in aiplatform.init will be used.

  • location (str) – Optional. Location to retrieve the Artifact from. If not set, locationset in aiplatform.init will be used.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to retrieve this Artifact. Overridescredentials set in aiplatform.init.

• Returns

  An ExperimentModel class that represents an Artifact resource.

• Raises

  ValueError – if artifact’s schema title is not ‘google.ExperimentModel’.

get_model_info()

Get the model’s info from an experiment model artifact.

• Returns

  A dict of model’s info. This includes model’s class name, framework name, framework version, and input example.

load_model()

Retrieves the original ML model from an ExperimentModel.

Example Usage:\experiment_model = aiplatform.get_experiment_model("my-sklearn-model")sk_model = experiment_model.load_model()pred_y = model.predict(test_X)``

• Returns

  The original ML model.

• Raises

  ValueError – if model type is not supported.

property model_class(: Optional[str )

The class name of the saved ML model.

register_model(*, model_id:Optional[str] = None, parent_model:Optional[str] = None, use_gpu:bool = False, is_default_version:bool = True, version_aliases:Optional[Sequence[str]] = None, version_description:Optional[str] = None, display_name:Optional[str] = None, description:Optional[str] = None, labels:Optional[Dict[str,str]] = None, serving_container_image_uri:Optional[str] = None, serving_container_predict_route:Optional[str] = None, serving_container_health_route:Optional[str] = None, serving_container_command:Optional[Sequence[str]] = None, serving_container_args:Optional[Sequence[str]] = None, serving_container_environment_variables:Optional[Dict[str,str]] = None, serving_container_ports:Optional[Sequence[int]] = None, instance_schema_uri:Optional[str] = None, parameters_schema_uri:Optional[str] = None, prediction_schema_uri:Optional[str] = None, explanation_metadata:Optional[google.cloud.aiplatform_v1.types.explanation_metadata.ExplanationMetadata] = None, explanation_parameters:Optional[google.cloud.aiplatform_v1.types.explanation.ExplanationParameters] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, encryption_spec_key_name:Optional[str] = None, staging_bucket:Optional[str] = None, sync:Optional[bool] = True, upload_request_timeout:Optional[float] = None)

Register an ExperimentModel to Model Registry and returns a Model representing the registered Model resource.

Example Usage:\experiment_model = aiplatform.get_experiment_model("my-sklearn-model")registered_model = experiment_model.register_model()registered_model.deploy(endpoint=my_endpoint)``

• Parameters

  • model_id (str) – Optional. The ID to use for the registered Model, which willbecome the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model that thenewly-registered model will be a version of.Only set this field when uploading a new version of an existing model.

  • use_gpu (str) – Optional. Whether or not to use GPUs for the serving container. Onlyspecify this argument when registering a Tensorflow model and‘serving_container_image_uri’ is not specified.

  • is_default_version (bool) – Optional. When set to True, the newly registered model version willautomatically have alias “default” included. Subsequent uses ofthis model without a version specified will use this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the newly-registered model version will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • version_aliases (Sequence[str]) – Optional. User provided version aliases so that a model versioncan be referenced via alias instead of auto-generated version ID.A default version alias will be created for the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • version_description (str) – Optional. The description of the model version being uploaded.

  • display_name (str) – Optional. The display name of the Model. The name can be up to 128characters long and can be consist of any UTF-8 characters.

  • description (str) – Optional. The description of the model.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • serving_container_image_uri (str) – Optional. The URI of the Model serving container. A pre-built container[https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers](https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers)is automatically chosen based on the model’s framwork. Set this field tooverride the default pre-built container.

  • serving_container_predict_route (str) – Optional. An HTTP path to send prediction requests to the container, andwhich must be supported by it. If not specified a default HTTP path willbe used by Vertex AI.

  • serving_container_health_route (str) – Optional. An HTTP path to send health check requests to the container, and whichmust be supported by it. If not specified a standard HTTP path will beused by Vertex AI.

  • serving_container_command (Sequence[str]) – Optional. The command with which the container is run. Not executed within ashell. The Docker image’s ENTRYPOINT is used if this is not provided.Variable references $(VAR_NAME) are expanded using the container’senvironment. If a variable cannot be resolved, the reference in theinput string will be unchanged. The $(VAR_NAME) syntax can be escapedwith a double $$, ie: $$(VAR_NAME). Escaped references will never beexpanded, regardless of whether the variable exists or not.

  • serving_container_args (Sequence[str]) – Optional. The arguments to the command. The Docker image’s CMD is used if this isnot provided. Variable references $(VAR_NAME) are expanded using thecontainer’s environment. If a variable cannot be resolved, the referencein the input string will be unchanged. The $(VAR_NAME) syntax can beescaped with a double $$, ie: $$(VAR_NAME). Escaped references willnever be expanded, regardless of whether the variable exists or not.

  • serving_container_environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The environment variables that are to be present in the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.

  • serving_container_ports (Sequence[int]) – Optional. Declaration of ports that are exposed by the container. This field isprimarily informational, it gives Vertex AI information about thenetwork connections the container uses. Listing or not a port here hasno impact on whether the port is actually exposed, any port listening onthe default “0.0.0.0” address inside a container will be accessible fromthe network.

  • instance_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the format of a single instance, whichare used inPredictRequest.instances,ExplainRequest.instancesandBatchPredictionJob.input_config.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform. Note: The URI given on output will be immutableand probably different, including the URI scheme, than theone given on input. The output URI will point to a locationwhere the user only has a read access.

  • parameters_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the parameters of prediction andexplanation viaPredictRequest.parameters,ExplainRequest.parametersandBatchPredictionJob.model_parameters.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform, if no parameters are supported it is set to anempty string. Note: The URI given on output will beimmutable and probably different, including the URI scheme,than the one given on input. The output URI will point to alocation where the user only has a read access.

  • prediction_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the format of a single predictionproduced by this Model, which are returned viaPredictResponse.predictions,ExplainResponse.explanations,andBatchPredictionJob.output_config.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform. Note: The URI given on output will be immutableand probably different, including the URI scheme, than theone given on input. The output URI will point to a locationwhere the user only has a read access.

  • explanation_metadata (aiplatform.explain.ExplanationMetadata) – Optional. Metadata describing the Model’s input and output for explanation.explanation_metadata is optional while explanation_parameters must bespecified when used.For more details, see Ref docshttp://tinyurl.com/1igh60kt

  • explanation_parameters (aiplatform.explain.ExplanationParameters) – Optional. Parameters to configure explaining for Model’s predictions.For more details, see Ref docshttp://tinyurl.com/1an4zake

  • project – Optional[str]=None,Project to upload this model to. Overrides project set inaiplatform.init.

  • location – Optional[str]=None,Location to upload this model to. Overrides location set inaiplatform.init.

  • credentials – Optional[auth_credentials.Credentials]=None,Custom credentials to use to upload this model. Overrides credentialsset in aiplatform.init.

  • encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theformprojects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, this Model and all sub-resources of this Model will be secured by this key.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • staging_bucket (str) – Optional. Bucket to stage local model artifacts. Overridesstaging_bucket set in aiplatform.init.

  • sync (bool) – Optional. Whether to execute this method synchronously. If False,this method will unblock and it will be executed in a concurrent Future.

  • upload_request_timeout (float) – Optional. The timeout for the upload request in seconds.

• Returns

  Instantiated representation of the registered model resource.

• Return type

  model (aiplatform.Model)

• Raises

  ValueError – If the model doesn’t have a pre-built container that is suitable for its framework and ‘serving_container_image_uri’ is not set.

class google.cloud.aiplatform.Artifact(artifact_name:str, *, metadata_store_id:str = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Bases:google.cloud.aiplatform.metadata.resource._Resource

Metadata Artifact resource for Vertex AI

Retrieves an existing Metadata Artifact given a resource name or ID.

• Parameters

  • artifact_name (str) – Required. A fully-qualified resource name or resource ID of the Artifact.Example: “projects/123/locations/us-central1/metadataStores/default/artifacts/my-resource”.or “my-resource” when project and location are initialized or passed.

  • metadata_store_id (str) – Optional. MetadataStore to retrieve Artifact from. If not set, metadata_store_id is set to “default”.If artifact_name is a fully-qualified resource, its metadata_store_id overrides this one.

  • project (str) – Optional. Project to retrieve the artifact from. If not set, projectset in aiplatform.init will be used.

  • location (str) – Optional. Location to retrieve the Artifact from. If not set, locationset in aiplatform.init will be used.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to retrieve this Artifact. Overridescredentials set in aiplatform.init.

client_class()

alias ofgoogle.cloud.aiplatform.utils.MetadataClientWithOverride

classmethod create(schema_title:str, *, resource_id:Optional[str] = None, uri:Optional[str] = None, display_name:Optional[str] = None, schema_version:Optional[str] = None, description:Optional[str] = None, metadata:Optional[Dict] = None, state:google.cloud.aiplatform_v1.types.artifact.Artifact.State = State.LIVE, metadata_store_id:Optional[str] = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Creates a new Metadata Artifact.

• Parameters

  • schema_title (str) – Required. schema_title identifies the schema title used by the Artifact.

    Please referencehttps://cloud.google.com/vertex-ai/docs/ml-metadata/system-schemas.

  • resource_id (str) – Optional. The <resource_id> portion of the Artifact name withthe format. This is globally unique in a metadataStore:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/artifacts/<resource_id>.

  • uri (str) – Optional. The uniform resource identifier of the artifact file. May be empty if there is no actualartifact file.

  • display_name (str) – Optional. The user-defined name of the Artifact.

  • schema_version (str) – Optional. schema_version specifies the version used by the Artifact.If not set, defaults to use the latest version.

  • description (str) – Optional. Describes the purpose of the Artifact to be created.

  • metadata (Dict) – Optional. Contains the metadata information that will be stored in the Artifact.

  • state (google.cloud.gapic.types.Artifact.State) – Optional. The state of this Artifact. This is aproperty of the Artifact, and does not imply orcapture any ongoing process. This property ismanaged by clients (such as Vertex AIPipelines), and the system does not prescribe orcheck the validity of state transitions.

  • metadata_store_id (str) – Optional. The <metadata_store_id> portion of the resource name withthe format:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/artifacts/<resource_id>If not provided, the MetadataStore’s ID will be set to “default”.

  • project (str) – Optional. Project used to create this Artifact. Overrides project set inaiplatform.init.

  • location (str) – Optional. Location used to create this Artifact. Overrides location set inaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials used to create this Artifact. Overridescredentials set in aiplatform.init.

• Returns

  Instantiated representation of the managed Metadata Artifact.

• Return type

  Artifact

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_id:str, metadata_store_id:str = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Retrieves a Metadata resource.

• Parameters

  • resource_id (str) – Required. The <resource_id> portion of the resource name with the format:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/<resource_noun>/<resource_id>.

  • metadata_store_id (str) – The <metadata_store_id> portion of the resource name withthe format:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/<resource_noun>/<resource_id>If not provided, the MetadataStore’s ID will be set to “default”.

  • project (str) – Project used to retrieve or create this resource. Overrides project set inaiplatform.init.

  • location (str) – Location used to retrieve or create this resource. Overrides location set inaiplatform.init.

  • credentials (auth_credentials.Credentials) – Custom credentials used to retrieve or create this resource. Overridescredentials set in aiplatform.init.

• Returns

  Instantiated representation of the managed Metadata resource or None if no resource was found.

• Return type

  resource (_Resource)

classmethod get_or_create(resource_id:str, schema_title:str, display_name:Optional[str] = None, schema_version:Optional[str] = None, description:Optional[str] = None, metadata:Optional[Dict] = None, metadata_store_id:str = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Retrieves or Creates (if it does not exist) a Metadata resource.

• Parameters

  • resource_id (str) – Required. The <resource_id> portion of the resource name with the format:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/<resource_noun>/<resource_id>.

  • schema_title (str) – Required. schema_title identifies the schema title used by the resource.

  • display_name (str) – Optional. The user-defined name of the resource.

  • schema_version (str) – Optional. schema_version specifies the version used by the resource.If not set, defaults to use the latest version.

  • description (str) – Optional. Describes the purpose of the resource to be created.

  • metadata (Dict) – Optional. Contains the metadata information that will be stored in the resource.

  • metadata_store_id (str) – The <metadata_store_id> portion of the resource name withthe format:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/<resource_noun>/<resource_id>If not provided, the MetadataStore’s ID will be set to “default”.

  • project (str) – Project used to retrieve or create this resource. Overrides project set inaiplatform.init.

  • location (str) – Location used to retrieve or create this resource. Overrides location set inaiplatform.init.

  • credentials (auth_credentials.Credentials) – Custom credentials used to retrieve or create this resource. Overridescredentials set in aiplatform.init.

• Returns

  Instantiated representation of the managed Metadata resource.

• Return type

  resource (_Resource)

classmethod get_with_uri(uri:str, *, metadata_store_id:Optional[str] = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Get an Artifact by it’s uri.

If more than one Artifact with this uri is in the metadata store then the Artifact with the latestcreate_time is returned.

• Parameters

  • uri (str) – Required. Uri of the Artifact to retrieve.

  • metadata_store_id (str) – Optional. MetadataStore to retrieve Artifact from. If not set, metadata_store_id is set to “default”.If artifact_name is a fully-qualified resource, its metadata_store_id overrides this one.

  • project (str) – Optional. Project to retrieve the artifact from. If not set, projectset in aiplatform.init will be used.

  • location (str) – Optional. Location to retrieve the Artifact from. If not set, locationset in aiplatform.init will be used.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to retrieve this Artifact. Overridescredentials set in aiplatform.init.

• Returns

  Artifact with given uri.

• Return type

  Artifact

• Raises

  ValueError – If no Artifact exists with the provided uri.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

property lineage_console_uri(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Cloud console uri to view this Artifact Lineage.

classmethod list(filter:Optional[str] = None, metadata_store_id:str = 'default', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, order_by:Optional[str] = None)

List resources that match the list filter in target metadataStore.

• Parameters

  • filter (str) – Optional. A query to filter available resources formatching results.

  • metadata_store_id (str) – The <metadata_store_id> portion of the resource name withthe format:projects/123/locations/us-central1/metadataStores/<metadata_store_id>/<resource_noun>/<resource_id>If not provided, the MetadataStore’s ID will be set to “default”.

  • project (str) – Project used to create this resource. Overrides project set inaiplatform.init.

  • location (str) – Location used to create this resource. Overrides location set inaiplatform.init.

  • credentials (auth_credentials.Credentials) – Custom credentials used to create this resource. Overridescredentials set in aiplatform.init.

  • order_by (str) – Optional. How the list of messages is ordered.Specify the values to order by and an ordering operation. Thedefault sorting order is ascending. To specify descending orderfor a field, users append a ” desc” suffix; for example: “foodesc, bar”. Subfields are specified with a. character, suchas foo.bar. seehttps://google.aip.dev/132#ordering for moredetails.

• Returns

  a list of managed Metadata resource.

• Return type

  resources (sequence[_Resource])

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

property state(: Optional[[google.cloud.aiplatform_v1.types.artifact.Artifact.State](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.Artifact.State)_ )

The State for this Artifact.

sync_resource()

Syncs local resource with the resource in metadata store.

to_dict()

Returns the resource proto as a dictionary.

update(metadata:Optional[Dict] = None, description:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, location:Optional[str] = None)

Updates an existing Metadata resource with new metadata.

• Parameters

  • metadata (Dict) – Optional. metadata contains the updated metadata information.

  • description (str) – Optional. Description describes the resource to be updated.

  • credentials (auth_credentials.Credentials) – Custom credentials to use to update this resource. Overridescredentials set in aiplatform.init.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

property uri(: Optional[str )

Uri for this Artifact.

wait()

Helper method that blocks until all futures are complete.

class google.cloud.aiplatform.AutoMLForecastingTrainingJob(display_name:Optional[str] = None, optimization_objective:Optional[str] = None, column_specs:Optional[Dict[str,str]] = None, column_transformations:Optional[List[Dict[str,Dict[str,str]]]] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._ForecastingTrainingJob

Class to train AutoML forecasting models.

The AutoMLForecastingTrainingJob class uses the AutoML training methodto train and run a forecasting model. The AutoML training method is a goodchoice for most forecasting use cases. If your use case doesn’t benefit fromthe Seq2seq or the Temporal fusion transformer training method offeredby theSequenceToSequencePlusForecastingTrainingJoband[TemporalFusionTransformerForecastingTrainingJob]https://cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform.TemporalFusionTransformerForecastingTrainingJob)classes respectively, then AutoML is likely the best training method foryour forecasting predictions.

For sample code that shows you how to use AutoMLForecastingTrainingJob seetheCreate a training pipeline forecasting sampleon GitHub.

Constructs a Forecasting Training Job.

• Parameters

  • display_name (str) – Optional. The user-defined name of this TrainingPipeline.

  • optimization_objective (str) – Optional. Objective function the model is to be optimized towards.The training process creates a Model that optimizes the value of the objectivefunction over the validation set. The supported optimization objectives:“minimize-rmse” (default) - Minimize root-mean-squared error (RMSE).“minimize-mae” - Minimize mean-absolute error (MAE).“minimize-rmsle” - Minimize root-mean-squared log error (RMSLE).“minimize-rmspe” - Minimize root-mean-squared percentage error (RMSPE).“minimize-wape-mae” - Minimize the combination of weighted absolute percentage error (WAPE)

    and mean-absolute-error (MAE).

    ”minimize-quantile-loss” - Minimize the quantile loss at the defined quantiles.

    (Set this objective to build quantile forecasts.)

  • column_specs (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. Alternative to column_transformations where the keys of the dictare column names and their respective values are one ofAutoMLTabularTrainingJob.column_data_types.When creating transformation for BigQuery Struct column, the columnshould be flattened using “.” as the delimiter. Only columns with no childshould have a transformation.If an input column has no transformations on it, such a column isignored by the training, except for the targetColumn, which should haveno transformations defined on.Only one of column_transformations or column_specs should be passed.

  • column_transformations (List[Dict[str, **Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]]]*) – Optional. Transformations to apply to the input columns (i.e. columns otherthan the targetColumn). Each transformation may produce multipleresult values from the column’s value, and all are used for training.When creating transformation for BigQuery Struct column, the columnshould be flattened using “.” as the delimiter. Only columns with no childshould have a transformation.If an input column has no transformations on it, such a column isignored by the training, except for the targetColumn, which should haveno transformations defined on.Only one of column_transformations or column_specs should be passed.Consider using column_specs as column_transformations will be deprecated eventually.

  • project (str) – Optional. Project to run training in. Overrides project set in aiplatform.init.

  • location (str) – Optional. Location to run training in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to run call training service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize TrainingPipelines.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the training pipeline. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.If set, this TrainingPipeline will be secured by this key.Note: Model trained by this TrainingPipeline is also securedby this key ifmodel_to_upload is not set separately.Overrides encryption_spec_key_name set in aiplatform.init.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.If set, the trained Model will be secured by this key.Overrides encryption_spec_key_name set in aiplatform.init.

• Raises

  ValueError – If both column_transformations and column_specs were provided.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property evaluated_data_items_bigquery_uri(: Optional[str )

BigQuery location of exported evaluated examples from the Training Job:returns:

BigQuery uri for the exported evaluated examples if the export

feature is enabled for training.

None: If the export feature was not enabled for training.

• Return type

  str

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset: google.cloud.aiplatform.datasets.time_series_dataset.TimeSeriesDataset, target_column:str, time_column:str, time_series_identifier_column:str, unavailable_at_forecast_columns:List[str], available_at_forecast_columns:List[str], forecast_horizon:int, data_granularity_unit:str, data_granularity_count:int, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, predefined_split_column_name:Optional[str] = None, timestamp_split_column_name:Optional[str] = None, weight_column:Optional[str] = None, time_series_attribute_columns:Optional[List[str]] = None, context_window:Optional[int] = None, export_evaluated_data_items:bool = False, export_evaluated_data_items_bigquery_destination_uri:Optional[str] = None, export_evaluated_data_items_override_destination:bool = False, quantiles:Optional[List[float]] = None, validation_options:Optional[str] = None, budget_milli_node_hours:int = 1000, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, additional_experiments:Optional[List[str]] = None, hierarchy_group_columns:Optional[List[str]] = None, hierarchy_group_total_weight:Optional[float] = None, hierarchy_temporal_total_weight:Optional[float] = None, hierarchy_group_temporal_total_weight:Optional[float] = None, window_column:Optional[str] = None, window_stride_length:Optional[int] = None, window_max_count:Optional[int] = None, holiday_regions:Optional[List[str]] = None, sync:bool = True, create_request_timeout:Optional[float] = None, enable_probabilistic_inference:bool = False)

Runs the training job and returns a model.

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:Any of `training_fraction_split`, `validation_fraction_split` and`test_fraction_split` may optionally be provided, they must sum to up to 1. Ifthe provided ones sum to less than 1, the remainder is assigned to sets asdecided by Vertex AI. If none of the fractions are set, by default roughly 80%of data will be used for training, 10% for validation, and 10% for test.Predefined splits:Assigns input data to training, validation, and test sets based on the value of a provided key.If using predefined splits, `predefined_split_column_name` must be provided.Supported only for tabular Datasets.Timestamp splits:Assigns input data to training, validation, and test setsbased on a provided timestamps. The youngest data pieces areassigned to training set, next to validation set, and the oldestto the test set.Supported only for tabular Datasets.

• Parameters

  • dataset (datasets.TimeSeriesDataset) – Required. The dataset within the same Project from which data will be used to train the Model. TheDataset must use schema compatible with Model being trained,and what is compatible should be described in the usedTrainingPipeline’s [training_task_definition][google.cloud.aiplatform.v1beta1.TrainingPipeline.training_task_definition].For time series Datasets, all their data is exported totraining, to pick and choose from.

  • target_column (str) – Required. Name of the column that the Model is to predict values for. Thiscolumn must be unavailable at forecast.

  • time_column (str) – Required. Name of the column that identifies time order in the time series.This column must be available at forecast.

  • time_series_identifier_column (str) – Required. Name of the column that identifies the time series.

  • unavailable_at_forecast_columns (List[str]) – Required. Column names of columns that are unavailable at forecast.Each column contains information for the given entity (identified by the[time_series_identifier_column]) that is unknown before the forecast(e.g. population of a city in a given year, or weather on a given day).

  • available_at_forecast_columns (List[str]) – Required. Column names of columns that are available at forecast.Each column contains information for the given entity (identified by the[time_series_identifier_column]) that is known at forecast.

  • forecast_horizon – (int):Required. The amount of time into the future for which forecasted values for the target arereturned. Expressed in number of units defined by the [data_granularity_unit] and[data_granularity_count] field. Inclusive.

  • data_granularity_unit (str) – Required. The data granularity unit. Accepted values areminute,hour,day,week,month,year.

  • data_granularity_count (int) – Required. The number of data granularity units between data points in the trainingdata. If [data_granularity_unit] is minute, can be 1, 5, 10, 15, or 30. For all othervalues of [data_granularity_unit], must be 1.

  • predefined_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key (either the label’s value orvalue in the column) must be one of {TRAIN,VALIDATE,TEST}, and it defines to which set thegiven piece of data is assigned. If for a piece of data thekey is not present or has an invalid value, that piece isignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timestamp_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key values of the key (the values inthe column) must be in RFC 3339 date-time format, wheretime-offset = “Z” (e.g. 1985-04-12T23:20:50.52Z). If for apiece of data the key is not present or has an invalid value,that piece is ignored by the pipeline.Supported only for tabular and time series Datasets.This parameter must be used with training_fraction_split,validation_fraction_split, and test_fraction_split.

  • weight_column (str) – Optional. Name of the column that should be used as the weight column.Higher values in this column give more importance to the rowduring Model training. The column must have numeric values between 0 and10000 inclusively, and 0 value means that the row is ignored.If the weight column field is not set, then all rows are assumed to haveequal weight of 1.

  • time_series_attribute_columns (List[str]) – Optional. Column names that should be used as attribute columns.Each column is constant within a time series.

  • context_window (int) – Optional. The amount of time into the past training and prediction data is used formodel training and prediction respectively. Expressed in number of units defined by the[data_granularity_unit] and [data_granularity_count] fields. When not provided uses thedefault value of 0 which means the model sets each series context window to be 0 (alsoknown as “cold start”). Inclusive.

  • export_evaluated_data_items (bool) – Whether to export the test set predictions to a BigQuery table.If False, then the export is not performed.

  • export_evaluated_data_items_bigquery_destination_uri (string) – Optional. URI of desired destination BigQuery table for exported test set predictions.

    Expected format:bq://<project_id>:<dataset_id>:<table>

    If not specified, then results are exported to the following auto-created BigQuerytable:<project_id>:export_evaluated_examples_<model_name>_<yyyy_MM_dd'T'HH_mm_ss_SSS'Z'>.evaluated_examples

    Applies only if [export_evaluated_data_items] is True.

  • export_evaluated_data_items_override_destination (bool) – Whether to override the contents of [export_evaluated_data_items_bigquery_destination_uri],if the table exists, for exported test set predictions. If False, and thetable exists, then the training job will fail.

    Applies only if [export_evaluated_data_items] is True and[export_evaluated_data_items_bigquery_destination_uri] is specified.

  • quantiles (List[float]) – Quantiles to use for the minimize-quantile-loss[AutoMLForecastingTrainingJob.optimization_objective]. This argument is required inthis case.

    Accepts up to 5 quantiles in the form of a double from 0 to 1, exclusive.Each quantile must be unique.

  • validation_options (str) – Validation options for the data validation component. The available options are:“fail-pipeline” - (default), will validate against the validation and fail the pipeline

    if it fails.

    ”ignore-validation” - ignore the results of the validation and continue the pipeline

  • budget_milli_node_hours (int) – Optional. The train budget of creating this Model, expressed in milli nodehours i.e. 1,000 value in this field means 1 node hour.The training cost of the model will not exceed this budget. The finalcost will be attempted to be close to the budget, though may end upbeing (even) noticeably smaller - at the backend’s discretion. Thisespecially may happen when further model training ceases to provideany improvements.If the budget is set to a value known to be insufficient to train aModel for the given training set, the training won’t be attempted andwill error.The minimum value is 1000 and the maximum is 72000.

  • model_display_name (str) – Optional. If the script produces a managed Vertex AI Model. The display name ofthe Model. The name can be up to 128 characters long and can be consistof any UTF-8 characters.

    If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • additional_experiments (List[str]) – Optional. Additional experiment flags for the time series forcasting training.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

  • hierarchy_group_columns (List[str]) – Optional. A list of time series attribute column names thatdefine the time series hierarchy. Only one level of hierarchy issupported, ex.region for a hierarchy of stores ordepartment for a hierarchy of products. If multiple columnsare specified, time series will be grouped by their combinedvalues, ex. (blue,large) forcolor andsize, upto 5 columns are accepted. If no group columns are specified,all time series are considered to be part of the same group.

  • hierarchy_group_total_weight (float) – Optional. The weight of the loss for predictions aggregated overtime series in the same hierarchy group.

  • hierarchy_temporal_total_weight (float) – Optional. The weight of the loss for predictions aggregated overthe horizon for a single time series.

  • hierarchy_group_temporal_total_weight (float) – Optional. The weight of the loss for predictions aggregated overboth the horizon and time series in the same hierarchy group.

  • window_column (str) – Optional. Name of the column that should be used to filter inputrows. The column should contain either booleans or stringbooleans; if the value of the row is True, generate a slidingwindow from that row.

  • window_stride_length (int) – Optional. Step length used to generate input examples. Everywindow_stride_length rows will be used to generate a slidingwindow.

  • window_max_count (int) – Optional. Number of rows that should be used to generate inputexamples. If the total row count is larger than this number, theinput data will be randomly sampled to hit the count.

  • holiday_regions (List[str]) – Optional. The geographical regions to use when creating holidayfeatures. This option is only allowed when data_granularity_unitisday. Acceptable values can come from any of the followinglevels:

    Top level: GLOBALSecond level: continental regions

    NA: North AmericaJAPAC: Japan and Asia PacificEMEA: Europe, the Middle East and AfricaLAC: Latin America and the Caribbean

    Third level: countries from ISO 3166-1 Country codes.

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • enable_probabilistic_inference (bool) – If probabilistic inference is enabled, the model will fit adistribution that captures the uncertainty of a prediction. Atinference time, the predictive distribution is used to make apoint prediction that minimizes the optimization objective. Forexample, the mean of a predictive distribution is the pointprediction that minimizes RMSE loss. If quantiles are specified,then the quantiles of the distribution are also returned. Theoptimization objective cannot be minimize-quantile-loss.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run or is waiting to run.

property start_time(: Optional[datetime.datetime )

Optional. The time when the training job first entered thePIPELINE_STATE_RUNNING state.

property state(: Optional[[google.cloud.aiplatform_v1.types.pipeline_state.PipelineState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.PipelineState)_ )

Current training state.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_resource_creation()

Waits until the resource has been created.

class google.cloud.aiplatform.AutoMLImageTrainingJob(display_name:Optional[str] = None, prediction_type:str = 'classification', multi_label:bool = False, model_type:str = 'CLOUD', base_model:Optional[google.cloud.aiplatform.models.Model] = None, incremental_train_base_model:Optional[google.cloud.aiplatform.models.Model] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None, checkpoint_name:Optional[str] = None, trainer_config:Optional[Dict[str,str]] = None, metric_spec:Optional[Dict[str,str]] = None, parameter_spec:Optional[Dict[str,Union[google.cloud.aiplatform.hyperparameter_tuning.DoubleParameterSpec, google.cloud.aiplatform.hyperparameter_tuning.IntegerParameterSpec, google.cloud.aiplatform.hyperparameter_tuning.CategoricalParameterSpec, google.cloud.aiplatform.hyperparameter_tuning.DiscreteParameterSpec]]] = None, search_algorithm:Optional[str] = None, measurement_selection:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._TrainingJob

Creates an AutoML image training job.

Use the AutoMLImageTrainingJob class to create, train, and return animage model. For more information about working with image data modelsin Vertex AI, seeImage data.

For an example of how to use the AutoMLImageTrainingJob class, see thetutorial in theAutoML imageclassificationnotebook on GitHub.

• Parameters

  • display_name (str) – Optional. The user-defined name of the trainingpipeline. The name must contain 128 or fewer UTF-8 characters.

  • prediction_type (str) – The type of prediction the model produces.Valid values are: classification to predict one of multipletarget values that are selected from each row. object_detectionto predict a value based on its relation to other values.object_detection is available only for columns that containsemantically numeric values. For example, columns might containnumeric values stored as an integer, float, or string.

  • multi_label (bool = False) – Required. If false, a single-label(multi-class) model istrained (i.e. assuming that for each image just up to oneannotation may be applicable). If true, a multi-label model istrained. If you pass in true, it’s assumed multipleannotations might apply to each image. Multi-label models aresupported only when prediction_type is set toclassification. If multi_label is true andprediction_type is object_detection, then then multi_labelis ignored. The default value is false.

  • model_type (str = "CLOUD") – Required.The type of model to create. The following are the validvalues:

    CLOUD - The default for image classification. This model typeis deigned to work in Google Cloud and can’t be exported.

    CLOUD_1 - This model type is deigned to work in Google Cloudand can’t be exported. This model type is expected to havehigher prediction accuracy than CLOUD.

    CLOUD_HIGH_ACCURACY_1 - The default for image object detection.This model type is deigned to work in Google Cloud and can’t beexported. It’s designed to have a higher latency and a higherprediction quality than other cloud model types.

    CLOUD_LOW_LATENCY_1 - This model type is deigned to work inGoogle Cloud and can’t be exported. It’s also designed to havelow latency, but might might lower prediction quality than othercloud model types.

    MOBILE_TF_LOW_LATENCY_1 - This model type is deigned to workin Google Cloud and can be exported as a TensorFlow or Core MLmodel for use on a mobile or edge device. It’s also designed tohave to have low latency, but might have lower predictionquality than other mobile models types.

    MOBILE_TF_VERSATILE_1 - This model type is deigned to work inGoogle Cloud and can be exported as a TensorFlow or Core MLmodel for use on a mobile or edge device.

    MOBILE_TF_HIGH_ACCURACY_1 - This model type is deigned to workin Google Cloud and can be exported as a TensorFlow or Core MLmodel for use on a mobile or edge device. It’s designed to havea higher latency, but should also have a higher predictionquality than other mobile model types.

    EFFICIENTNET - This model type is available in the VertexModel Garden image classification training with customizablehyperparameters. It’s deigned to work in Google Cloud and can’tbe exported.

    VIT - This model type is available in the Vertex ModelGarden image classification training with customizablehyperparameters. Best tailored to be usedwithin Google Cloud, and cannot be exported externally.

    MAXVIT - This model type is available in the Vertex ModelGarden image classification training with customizablehyperparameters. It’s deigned to work in Google Cloud and can’tbe exported.

    COCA - This model type is available in the Vertex Model Gardenimage classification training with customizable hyperparameters.It’s deigned to work in Google Cloud and can’t be exported.

    SPINENET - This model type is available in the Vertex ModelGarden image object detection training with customizablehyperparameters. It’s deigned to work in Google Cloud and can’tbe exported.

    YOLO - This model type is available in the Vertex Model Gardenimage object detection training with customizablehyperparameters. It’s deigned to work in Google Cloud and can’tbe exported.

  • base_model – Optional[models.Model] = Optional. You can specify abase_model for image classification models only. If it isspecified, the new model is trained based on the base_model.Otherwise, the new model is trained from scratch. The base modelmust be in the same Google Project and region as the new model totrain, and it must have the same model_type.

  • incremental_train_base_model – Optional[models.Model] = Optional. Youcan specify an incremental_train_base_model for imageclassification models and object detection models only. Ifspecified, the new model is incrementally trained using anexisting model as the starting point. This can reduce thetraining time. If not specified, the new model is trained fromscratch. The model specified in incremental_train_base_modelmodel must be in the same Google Project and region as the newmodel to train, and it must have the same prediction_type andmodel_type.

  • project (str) – Optional. The Google Cloud region where this wherethe training runs. This region overrides the region that was setby aiplatform.init.

  • location (str) – Optional. Location to run training in. Overrideslocation set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Thecredentials that are used to train the model. These credentialsoverride the credentials set by aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. Labels with user-defined metadatato organize your training pipelines. The maximum length of a keyand value is 64 unicode characters. Labels and keys can containonly lowercase letters, numeric characters, underscores, anddashes. International characters are allowed. For moreinformation and examples of using labels, seeUsing labels toorganize Google Cloud Platformresources.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The CloudKMS resource identifier of the customer managed encryption keyused to protect the training pipeline. The key has the followingformat:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created. If set, the key secures the trainingpipeline and overrides the key set using aiplatform.init. Themodel trained by this training pipeline is also secured by thiskey if model_encryption_spec_key_name is not set separately.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The CloudKMS resource identifier of the customer managed encryption keyused to protect the model. The key has the following format:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created. If set, the key secures the trained modeland overrides the key set using aiplatform.init.

  • checkpoint_name – Optional[str] = Optional. The field is reserved forModel Garden model training and is based on the providedpre-trained model checkpoint. checkpoint_name is needed onlywhen model_type is set to EFFICIENTNET, VIT, COCA,SPINENET, or YOLO.

  • trainer_config – Optional[Dict[str, str]] = None, Optional. A fieldthat’s used with the Model Garden model training when passingcustomized configs for the trainer. The following is an examplethat uses all trainer_config parameters:

```python````````python````pytrainer_config = {> ’global_batch_size’: ‘8’,> ‘learning_rate’: ‘0.001’,> ‘optimizer_type’: ‘sgd’,> ‘optimizer_momentum’: ‘0.9’,> ‘train_steps’: ‘10000’,> ‘accelerator_count’: ‘1’,> ‘anchor_size’: ‘8’,  – IOD onlytrainer_config is required for only Model Garden models whenmodel_type is EFFICIENTNET, VIT, COCA, SPINENET, orYOLO.* **metric_spec** – Dict[str, str] Required. A dictionary that representsmetrics use for optimization. The dictionary key is themetric_id that’s reported by your training job and can be‘loss’ or ‘accuracy’. The dictionary value is the optimizationgoal of the metric, and can be ‘minimize’ or ‘maximize’. Thefollowing is an example of a metric_spec: metric_spec ={‘loss’: ‘minimize’, ‘accuracy’: ‘maximize’}. metric_spec isrequired for only Model Garden models when model_type isEFFICIENTNET, VIT, COCA, SPINENET, or YOLO.* **parameter_spec** (*Dict**[*[*str*](https://docs.python.org/3/library/stdtypes.html#str)*, **hpt._ParameterSpec**]*) – Required. Adictionary representing parameters to use for optimization.The dictionary key is the metric_id that’s reported by yourtraining job as a command line keyword argument. The dictionaryvalue is the parameter specification of the metric. Thefollowing is an example of how to specify a parameter_spec:```python````````python````pyfrom google.cloud.aiplatform import hpt as hptparameter_spec = {    ‘learning_rate’: hpt.DoubleParameterSpec(min=1e-7, max=1, scale=’linear’),Supported parameter specifications can be found inaiplatform.hyperparameter_tuning. The following parameterspecifications are supported: Union[DoubleParameterSpec,IntegerParameterSpec, CategoricalParameterSpec,DiscreteParameterSpec]. parameter_spec \`is required for onlyModel Garden models when \`model_type is EFFICIENTNET, VIT,COCA, SPINENET, or YOLO.* **search_algorithm** ([*str*](https://docs.python.org/3/library/stdtypes.html#str)) – The search algorithm that’s specified forthe study. The search algorithm can be one of the following:None - If you don’t specify an algorithm, your job uses thedefault Vertex AI algorithm. The default algorithm appliesBayesian optimization to arrive at the optimal solution with amore effective search over the parameter space.’grid’ - A simple grid search within the feasible space. Thisoption is particularly useful if you want to specify a quantityof trials that is greater than the number of points in thefeasible space. In these cases, if you don’t specify a gridsearch, the Vertex AI default algorithm might generate duplicatesuggestions. To use grid search, all parameter specificationsmust be IntegerParameterSpec, CategoricalParameterSpec, orDiscreteParameterSpec.’random’ - A simple random search within the feasible space.search_algorithm is required for only Model Garden models whenmodel_type is EFFICIENTNE\`T, VIT, COCA, SPINENET, orYOLO.* **measurement_selection** ([*str*](https://docs.python.org/3/library/stdtypes.html#str)) – Indicates which measurement to usewhen the service selects the final measurement from previouslyreported intermediate measurements. measurement_selection canbe best or last. If you expect your measurements tomonotonically improve, then choose last. If your system canover-train and you expect the performance to improve and thenstart to decline, then choose best. If your measurements aresignificantly noisy or not reproducible, then best is likelyto be over-optimistic, so in this case last is the preferredoption. measurement_selection is required for only ModelGarden models when model_type is EFFICIENTNET, VIT,COCA, SPINENET, or YOLO.

• Raises

  ValueError – When an invalid prediction_type or model_type is provided.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset: google.cloud.aiplatform.datasets.image_dataset.ImageDataset, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, training_filter_split:Optional[str] = None, validation_filter_split:Optional[str] = None, test_filter_split:Optional[str] = None, budget_milli_node_hours:Optional[int] = None, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, disable_early_stopping:bool = False, sync:bool = True, create_request_timeout:Optional[float] = None)

Runs the AutoML Image training job and returns a model.

If training on a Vertex AI dataset, you can use one of the following splitconfigurations:

Data fraction splits:Any oftraining_fraction_split,validation_fraction_split andtest_fraction_split may optionally be provided, they must sum to upto 1. Ifthe provided ones sum to less than 1, the remainder is assigned to setsasdecided by Vertex AI. If none of the fractions are set, by defaultroughly 80%of data will be used for training, 10% for validation, and 10% for test.

Data filter splits:Assigns input data to training, validation, and test setsbased on the given filters, data pieces not matched by anyfilter are ignored. Currently only supported for Datasetscontaining DataItems.If any of the filters in this message are to match nothing, thenthey can be set as ‘-’ (the minus sign).If using filter splits, all oftraining_filter_split,validation_filter_split andtest_filter_split must be provided.Supported only for unstructured Datasets.

• Parameters

  • dataset (datasets.ImageDataset) – Required. The dataset within the sameProject from which data will be used to train the Model. The Datasetmust use schema compatible with Model being trained, and what iscompatible should be described in the used TrainingPipeline’s[training_task_definition][google.cloud.aiplatform.v1beta1.TrainingPipeline.training_task_definition].For tabular Datasets, all their data is exported to training, to pickand choose from.

  • training_fraction_split (float) – Optional. The fraction of the inputdata that is to be used to train the Model. This is ignored if Datasetis not provided.

  • validation_fraction_split (float) – Optional. The fraction of the inputdata that is to be used to validate the Model. This is ignored ifDataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input datathat is to be used to evaluate the Model. This is ignored if Datasetis not provided.

  • training_filter_split (str) – Optional. A filter on DataItems of theDataset. DataItems that match this filter are used to train the Model.A filter with same syntax as the one used inDatasetService.ListDataItems may be used. If a single DataItem ismatched by more than one of the FilterSplit filters, then it isassigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • validation_filter_split (str) – Optional. A filter on DataItems of theDataset. DataItems that match this filter are used to validate theModel. A filter with same syntax as the one used inDatasetService.ListDataItems may be used. If a single DataItem ismatched by more than one of the FilterSplit filters, then it isassigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • test_filter_split (str) – Optional. A filter on DataItems of the Dataset.DataItems that match this filter are used to test the Model. A filterwith same syntax as the one used in DatasetService.ListDataItems maybe used. If a single DataItem is matched by more than one of theFilterSplit filters, then it is assigned to the first set that appliesto it in the training, validation, test order. This is ignored ifDataset is not provided.

  • budget_milli_node_hours (int) – Optional. The train budget of creatingthis Model, expressed in milli node hours i.e. 1,000 value in thisfield means 1 node hour. Defaults by prediction_type:classification - For Cloud models the budget must be: 8,000 -800,000 milli node hours (inclusive). The default value is 192,000which represents one day in wall time, assuming 8 nodes are used.object_detection - For Cloud models the budget must be: 20,000 -900,000 milli node hours (inclusive). The default value is 216,000which represents one day in wall time, assuming 9 nodes are used. Thetraining cost of the model will not exceed this budget. The final costwill be attempted to be close to the budget, though may end up being(even) noticeably smaller - at the backend’s discretion. Thisespecially may happen when further model training ceases to provideany improvements. If the budget is set to a value known to beinsufficient to train a Model for the given training set, the trainingwon’t be attempted and will error.

  • model_display_name (str) – Optional. The display name of the managedVertex AI Model. The name can be up to 128 characters long and can beconsist of any UTF-8 characters. If not provided upon creation, thejob’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-definedmetadata to organize your Models. Label keys and values can be nolonger than 64 characters (Unicode codepoints), can only containlowercase letters, numeric characters, underscores and dashes.International characters are allowed. Seehttps://goo.gl/xmQnxf formore information and examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by thisjob, which will become the final component of the model resource name.This value may be up to 63 characters, and valid characters are[a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of anexisting model. The new model uploaded by this job will be a versionof parent_model. Only set this field when training a new version ofan existing model.

  • is_default_version (bool) – Optional. When set to True, the newlyuploaded model version will automatically have alias “default”included. Subsequent uses of the model produced by this job without aversion specified will use this “default” version. When set to False,the “default” alias will not be moved. Actions targeting the modelversion produced by this job will need to specifically reference thisversion by ID or alias. New model uploads, i.e. version 1, willalways be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided versionaliases so that the model version uploaded by this job can bereferenced via alias instead of auto-generated version ID. A defaultversion alias will be created for the first version of the model. Theformat is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the modelversion being uploaded by this job.

  • disable_early_stopping – bool = False Required. If true, the entirebudget is used. This disables the early stopping feature. By default,the early stopping feature is enabled, which means that training mightstop before the entire training budget has been used, if furthertraining does no longer brings significant improvement to the model.

  • sync – bool = True Whether to execute this method synchronously. IfFalse, this method will be executed in concurrent Future and anydownstream object will be immediately returned and synced when theFuture has completed.

  • create_request_timeout (float) – Optional. The timeout for the createrequest in seconds.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run or is waiting to run.

property start_time(: Optional[datetime.datetime )

Optional. The time when the training job first entered thePIPELINE_STATE_RUNNING state.

property state(: Optional[[google.cloud.aiplatform_v1.types.pipeline_state.PipelineState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.PipelineState)_ )

Current training state.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_resource_creation()

Waits until the resource has been created.

class google.cloud.aiplatform.AutoMLTabularTrainingJob(display_name:str, optimization_prediction_type:str, optimization_objective:Optional[str] = None, column_specs:Optional[Dict[str,str]] = None, column_transformations:Optional[List[Dict[str,Dict[str,str]]]] = None, optimization_objective_recall_value:Optional[float] = None, optimization_objective_precision_value:Optional[float] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._TrainingJob

Constructs a AutoML Tabular Training Job.

Example usage:

job = training_jobs.AutoMLTabularTrainingJob(

display_name=”my_display_name”,optimization_prediction_type=”classification”,optimization_objective=”minimize-log-loss”,column_specs={“column_1”: “auto”, “column_2”: “numeric”},labels={‘key’: ‘value’},

)

• Parameters

  • display_name (str) – Required. The user-defined name of this TrainingPipeline.

  • optimization_prediction_type (str) – The type of prediction the Model is to produce.“classification” - Predict one out of multiple target values ispicked for each row.“regression” - Predict a value based on its relation to other values.This type is available only to columns that containsemantically numeric values, i.e. integers or floatingpoint number, even if stored as e.g. strings.

  • optimization_objective (str) – Optional. Objective function the Model is to be optimized towards. The trainingtask creates a Model that maximizes/minimizes the value of the objectivefunction over the validation set.

    The supported optimization objectives depend on the prediction type, andin the case of classification also the number of distinct values in thetarget column (two distint values -> binary, 3 or more distinct values-> multi class).If the field is not set, the default objective function is used.

    Classification (binary):“maximize-au-roc” (default) - Maximize the area under the receiver

    operating characteristic (ROC) curve.

    ”minimize-log-loss” - Minimize log loss.“maximize-au-prc” - Maximize the area under the precision-recall curve.“maximize-precision-at-recall” - Maximize precision for a specified

    recall value.

    ”maximize-recall-at-precision” - Maximize recall for a specified

    precision value.

    Classification (multi class):“minimize-log-loss” (default) - Minimize log loss.

    Regression:“minimize-rmse” (default) - Minimize root-mean-squared error (RMSE).“minimize-mae” - Minimize mean-absolute error (MAE).“minimize-rmsle” - Minimize root-mean-squared log error (RMSLE).

  • column_specs (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. Alternative to column_transformations where the keys of the dictare column names and their respective values are one ofAutoMLTabularTrainingJob.column_data_types.When creating transformation for BigQuery Struct column, the columnshould be flattened using “.” as the delimiter. Only columns with no childshould have a transformation.If an input column has no transformations on it, such a column isignored by the training, except for the targetColumn, which should haveno transformations defined on.Only one of column_transformations or column_specs should be passed. If noneof column_transformations or column_specs is passed, the local credentialsbeing used will try setting column_specs to “auto”. To do this, the localcredentials require read access to the GCS or BigQuery training data source.

  • column_transformations (List[Dict[str, **Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]]]*) – Optional. Transformations to apply to the input columns (i.e. columns otherthan the targetColumn). Each transformation may produce multipleresult values from the column’s value, and all are used for training.When creating transformation for BigQuery Struct column, the columnshould be flattened using “.” as the delimiter. Only columns with no childshould have a transformation.If an input column has no transformations on it, such a column isignored by the training, except for the targetColumn, which should haveno transformations defined on.Only one of column_transformations or column_specs should be passed.Consider using column_specs as column_transformations will be deprecatedeventually. If none of column_transformations or column_specs is passed,the local credentials being used will try setting column_transformations to“auto”. To do this, the local credentials require read access to the GCS orBigQuery training data source.

  • optimization_objective_recall_value (float) – Optional. Required when maximize-precision-at-recall optimizationObjective waspicked, represents the recall value at which the optimization is done.

    The minimum value is 0 and the maximum is 1.0.

  • optimization_objective_precision_value (float) – Optional. Required when maximize-recall-at-precision optimizationObjective waspicked, represents the precision value at which the optimization isdone.

    The minimum value is 0 and the maximum is 1.0.

  • project (str) – Optional. Project to run training in. Overrides project set in aiplatform.init.

  • location (str) – Optional. Location to run training in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to run call training service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize TrainingPipelines.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the training pipeline. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, this TrainingPipeline will be secured by this key.

    Note: Model trained by this TrainingPipeline is also securedby this key ifmodel_to_upload is not set separately.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, the trained Model will be secured by this key.

    Overrides encryption_spec_key_name set in aiplatform.init.

• Raises

  ValueError – If both column_transformations and column_specs were provided.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

static get_auto_column_specs(dataset: google.cloud.aiplatform.datasets.tabular_dataset.TabularDataset, target_column:str)

Returns a dict with all non-target columns as keys and ‘auto’ as values.

Example usage:

column_specs = training_jobs.AutoMLTabularTrainingJob.get_auto_column_specs(

dataset=my_dataset,target_column=”my_target_column”,

)

• Parameters

  • dataset (datasets.TabularDataset) – Required. Intended dataset.

  • target_column (str) – Required. Intended target column.

• Returns

  Dict[str, str]

    Column names as keys and ‘auto’ as values

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset: google.cloud.aiplatform.datasets.tabular_dataset.TabularDataset, target_column:str, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, predefined_split_column_name:Optional[str] = None, timestamp_split_column_name:Optional[str] = None, weight_column:Optional[str] = None, budget_milli_node_hours:int = 1000, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, disable_early_stopping:bool = False, export_evaluated_data_items:bool = False, export_evaluated_data_items_bigquery_destination_uri:Optional[str] = None, export_evaluated_data_items_override_destination:bool = False, additional_experiments:Optional[List[str]] = None, sync:bool = True, create_request_timeout:Optional[float] = None)

Runs the training job and returns a model.

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:Any of `training_fraction_split`, `validation_fraction_split` and`test_fraction_split` may optionally be provided, they must sum to up to 1. Ifthe provided ones sum to less than 1, the remainder is assigned to sets asdecided by Vertex AI. If none of the fractions are set, by default roughly 80%of data will be used for training, 10% for validation, and 10% for test.Predefined splits:Assigns input data to training, validation, and test sets based on the value of a provided key.If using predefined splits, `predefined_split_column_name` must be provided.Supported only for tabular Datasets.Timestamp splits:Assigns input data to training, validation, and test setsbased on a provided timestamps. The youngest data pieces areassigned to training set, next to validation set, and the oldestto the test set.Supported only for tabular Datasets.

• Parameters

  • dataset (datasets.TabularDataset) – Required. The dataset within the same Project from which data will be used to train the Model. TheDataset must use schema compatible with Model being trained,and what is compatible should be described in the usedTrainingPipeline’s [training_task_definition][google.cloud.aiplatform.v1beta1.TrainingPipeline.training_task_definition].For tabular Datasets, all their data is exported totraining, to pick and choose from.

  • target_column (str) – Required. The name of the column values of which the Model is to predict.

  • training_fraction_split (float) – Optional. The fraction of the input data that is to be used to trainthe Model. This is ignored if Dataset is not provided.

  • validation_fraction_split (float) – Optional. The fraction of the input data that is to be used to validatethe Model. This is ignored if Dataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input data that is to be used to evaluatethe Model. This is ignored if Dataset is not provided.

  • predefined_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key (either the label’s value orvalue in the column) must be one of {training,validation,test}, and it defines to which set thegiven piece of data is assigned. If for a piece of data thekey is not present or has an invalid value, that piece isignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timestamp_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key values of the key (the values inthe column) must be in RFC 3339 date-time format, wheretime-offset = “Z” (e.g. 1985-04-12T23:20:50.52Z). If for apiece of data the key is not present or has an invalid value,that piece is ignored by the pipeline.Supported only for tabular and time series Datasets.This parameter must be used with training_fraction_split,validation_fraction_split, and test_fraction_split.

  • weight_column (str) – Optional. Name of the column that should be used as the weight column.Higher values in this column give more importance to the rowduring Model training. The column must have numeric values between 0 and10000 inclusively, and 0 value means that the row is ignored.If the weight column field is not set, then all rows are assumed to haveequal weight of 1.

  • budget_milli_node_hours (int) – Optional. The train budget of creating this Model, expressed in milli nodehours i.e. 1,000 value in this field means 1 node hour.The training cost of the model will not exceed this budget. The finalcost will be attempted to be close to the budget, though may end upbeing (even) noticeably smaller - at the backend’s discretion. Thisespecially may happen when further model training ceases to provideany improvements.If the budget is set to a value known to be insufficient to train aModel for the given training set, the training won’t be attempted andwill error.The minimum value is 1000 and the maximum is 72000.

  • model_display_name (str) – Optional. If the script produces a managed Vertex AI Model. The display name ofthe Model. The name can be up to 128 characters long and can be consistof any UTF-8 characters.

    If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • disable_early_stopping (bool) – Required. If true, the entire budget is used. This disables the early stoppingfeature. By default, the early stopping feature is enabled, which meansthat training might stop before the entire training budget has beenused, if further training does no longer brings significant improvementto the model.

  • export_evaluated_data_items (bool) – Whether to export the test set predictions to a BigQuery table.If False, then the export is not performed.

  • export_evaluated_data_items_bigquery_destination_uri (string) – Optional. URI of desired destination BigQuery table for exported test set predictions.

    Expected format:bq://<project_id>:<dataset_id>:<table>

    If not specified, then results are exported to the following auto-created BigQuerytable:<project_id>:export_evaluated_examples_<model_name>_<yyyy_MM_dd'T'HH_mm_ss_SSS'Z'>.evaluated_examples

    Applies only if [export_evaluated_data_items] is True.

  • export_evaluated_data_items_override_destination (bool) – Whether to override the contents of [export_evaluated_data_items_bigquery_destination_uri],if the table exists, for exported test set predictions. If False, and thetable exists, then the training job will fail.

    Applies only if [export_evaluated_data_items] is True and[export_evaluated_data_items_bigquery_destination_uri] is specified.

  • additional_experiments (List[str]) – Optional. Additional experiment flags for the automl tables training.

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run or is waiting to run.

property start_time(: Optional[datetime.datetime )

Optional. The time when the training job first entered thePIPELINE_STATE_RUNNING state.

property state(: Optional[[google.cloud.aiplatform_v1.types.pipeline_state.PipelineState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.PipelineState)_ )

Current training state.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_resource_creation()

Waits until the resource has been created.

class google.cloud.aiplatform.AutoMLTextTrainingJob(display_name:str, prediction_type:str, multi_label:bool = False, sentiment_max:int = 10, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._TrainingJob

Constructs a AutoML Text Training Job.

• Parameters

  • display_name (str) – Required. The user-defined name of this TrainingPipeline.

  • prediction_type (str) – The type of prediction the Model is to produce, one of:

    ”classification” - A classification model analyzes text data and

      returns a list of categories that apply to the text found in the data.  Vertex AI offers both single-label and multi-label text classification models.

    ”extraction” - An entity extraction model inspects text data

      for known entities referenced in the data and  labels those entities in the text.

    ”sentiment” - A sentiment analysis model inspects text data and identifies the

      prevailing emotional opinion within it, especially to determine a writer’s attitude  as positive, negative, or neutral.

  • multi_label (bool) – Required and only applicable for text classification task. If false, a single-label (multi-class) Model will be trained (i.e.assuming that for each text snippet just up to one annotation may beapplicable). If true, a multi-label Model will be trained (i.e.assuming that for each text snippet multiple annotations may beapplicable).

  • sentiment_max (int) – Required and only applicable for sentiment task. A sentiment is expressed as an integerordinal, where higher value means a morepositive sentiment. The range of sentiments thatwill be used is between 0 and sentimentMax(inclusive on both ends), and all the values inthe range must be represented in the datasetbefore a model can be created.Only the Annotations with this sentimentMax willbe used for training. sentimentMax value must bebetween 1 and 10 (inclusive).

  • project (str) – Optional. Project to run training in. Overrides project set in aiplatform.init.

  • location (str) – Optional. Location to run training in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to run call training service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize TrainingPipelines.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the training pipeline. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, this TrainingPipeline will be secured by this key.

    Note: Model trained by this TrainingPipeline is also securedby this key ifmodel_to_upload is not set separately.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, the trained Model will be secured by this key.

    Overrides encryption_spec_key_name set in aiplatform.init.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset: google.cloud.aiplatform.datasets.text_dataset.TextDataset, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, training_filter_split:Optional[str] = None, validation_filter_split:Optional[str] = None, test_filter_split:Optional[str] = None, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, sync:bool = True, create_request_timeout:Optional[float] = None)

Runs the training job and returns a model.

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:Any of `training_fraction_split`, `validation_fraction_split` and`test_fraction_split` may optionally be provided, they must sum to up to 1. Ifthe provided ones sum to less than 1, the remainder is assigned to sets asdecided by Vertex AI. If none of the fractions are set, by default roughly 80%of data will be used for training, 10% for validation, and 10% for test.Data filter splits:Assigns input data to training, validation, and test setsbased on the given filters, data pieces not matched by anyfilter are ignored. Currently only supported for Datasetscontaining DataItems.If any of the filters in this message are to match nothing, thenthey can be set as ‘-’ (the minus sign).If using filter splits, all of `training_filter_split`, `validation_filter_split` and`test_filter_split` must be provided.Supported only for unstructured Datasets.

• Parameters

  • dataset (datasets.TextDataset) – Required. The dataset within the same Project from which data will be used to train the Model. TheDataset must use schema compatible with Model being trained,and what is compatible should be described in the usedTrainingPipeline’s [training_task_definition][google.cloud.aiplatform.v1beta1.TrainingPipeline.training_task_definition].

  • training_fraction_split (float) – Optional. The fraction of the input data that is to be used to trainthe Model. This is ignored if Dataset is not provided.

  • validation_fraction_split (float) – Optional. The fraction of the input data that is to be used to validatethe Model. This is ignored if Dataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input data that is to be used to evaluatethe Model. This is ignored if Dataset is not provided.

  • training_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to train the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • validation_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to validate the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • test_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to test the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • model_display_name (str) – Optional. The display name of the managed Vertex AI Model.The name can be up to 128 characters long and can consistof any UTF-8 characters.

    If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels..

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds

• Returns

  The trained Vertex AI Model resource.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run or is waiting to run.

property start_time(: Optional[datetime.datetime )

Optional. The time when the training job first entered thePIPELINE_STATE_RUNNING state.

property state(: Optional[[google.cloud.aiplatform_v1.types.pipeline_state.PipelineState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.PipelineState)_ )

Current training state.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_resource_creation()

Waits until the resource has been created.

class google.cloud.aiplatform.AutoMLVideoTrainingJob(display_name:Optional[str] = None, prediction_type:str = 'classification', model_type:str = 'CLOUD', project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._TrainingJob

Constructs a AutoML Video Training Job.

• Parameters

  • display_name (str) – Required. The user-defined name of this TrainingPipeline.

  • prediction_type (str) – The type of prediction the Model is to produce, one of:

    ”classification” - A video classification model classifies shots

      and segments in your videos according to your own defined labels.

    ”object_tracking” - A video object tracking model detects and tracks

      multiple objects in shots and segments. You can use these  models to track objects in your videos according to your  own pre-defined, custom labels.

    ”action_recognition” - A video action recognition model pinpoints

      the location of actions with short temporal durations (~1 second).

  • model_type – str = “CLOUD”Required. One of the following:

    ”CLOUD” - available for “classification”, “object_tracking” and “action_recognition”

    A Model best tailored to be used within Google Cloud,and which cannot be exported.

    ”MOBILE_VERSATILE_1” - available for “classification”, “object_tracking” and “action_recognition”

    A model that, in addition to being available within GoogleCloud, can also be exported (see ModelService.ExportModel)as a TensorFlow or TensorFlow Lite model and used on amobile or edge device with afterwards.

    ”MOBILE_CORAL_VERSATILE_1” - available only for “object_tracking”

    A versatile model that is meant to be exported (seeModelService.ExportModel) and used on a Google Coral device.

    ”MOBILE_CORAL_LOW_LATENCY_1” - available only for “object_tracking”

    A model that trades off quality for low latency, to beexported (see ModelService.ExportModel) and used on aGoogle Coral device.

    ”MOBILE_JETSON_VERSATILE_1” - available only for “object_tracking”

    A versatile model that is meant to be exported (seeModelService.ExportModel) and used on an NVIDIA Jetson device.

    ”MOBILE_JETSON_LOW_LATENCY_1” - available only for “object_tracking”

    A model that trades off quality for low latency, to beexported (see ModelService.ExportModel) and used on anNVIDIA Jetson device.

  • project (str) – Optional. Project to run training in. Overrides project set in aiplatform.init.

  • location (str) – Optional. Location to run training in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to run call training service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize TrainingPipelines.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the training pipeline. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, this TrainingPipeline will be secured by this key.

    Note: Model trained by this TrainingPipeline is also securedby this key ifmodel_to_upload is not set separately.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, the trained Model will be secured by this key.

    Overrides encryption_spec_key_name set in aiplatform.init.

• Raises

  ValueError – When an invalid prediction_type and/or model_type is provided.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset: google.cloud.aiplatform.datasets.video_dataset.VideoDataset, training_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, training_filter_split:Optional[str] = None, test_filter_split:Optional[str] = None, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, sync:bool = True, create_request_timeout:Optional[float] = None)

Runs the AutoML Video training job and returns a model.

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:`training_fraction_split`, and `test_fraction_split` may optionallybe provided, they must sum to up to 1. If none of the fractions are set,by default roughly 80% of data will be used for training, and 20% for test.Data filter splits:Assigns input data to training, validation, and test setsbased on the given filters, data pieces not matched by anyfilter are ignored. Currently only supported for Datasetscontaining DataItems.If any of the filters in this message are to match nothing, thenthey can be set as ‘-’ (the minus sign).If using filter splits, all of `training_filter_split`, `validation_filter_split` and`test_filter_split` must be provided.Supported only for unstructured Datasets.

• Parameters

  • dataset (datasets.VideoDataset) – Required. The dataset within the same Project from which data will be used to train the Model. TheDataset must use schema compatible with Model being trained,and what is compatible should be described in the usedTrainingPipeline’s [training_task_definition][google.cloud.aiplatform.v1beta1.TrainingPipeline.training_task_definition].For tabular Datasets, all their data is exported totraining, to pick and choose from.

  • training_fraction_split (float) – Optional. The fraction of the input data that is to be used to trainthe Model. This is ignored if Dataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input data that is to be used to evaluatethe Model. This is ignored if Dataset is not provided.

  • training_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to train the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • test_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to test the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • model_display_name (str) – Optional. The display name of the managed Vertex AI Model. The namecan be up to 128 characters long and can be consist of any UTF-8characters. If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • sync – bool = TrueWhether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run or is waiting to run.

property start_time(: Optional[datetime.datetime )

Optional. The time when the training job first entered thePIPELINE_STATE_RUNNING state.

property state(: Optional[[google.cloud.aiplatform_v1.types.pipeline_state.PipelineState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.PipelineState)_ )

Current training state.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_resource_creation()

Waits until the resource has been created.

class google.cloud.aiplatform.BatchPredictionJob(batch_prediction_job_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Bases:google.cloud.aiplatform.jobs._Job,google.cloud.aiplatform.base.PreviewMixin

Retrieves a BatchPredictionJob resource and instantiates itsrepresentation.

• Parameters

  • batch_prediction_job_name (str) – Required. A fully-qualified BatchPredictionJob resource name or ID.Example: “projects/…/locations/…/batchPredictionJobs/456” or“456” when project and location are initialized or passed.

  • project – Optional[str] = None,Optional. project to retrieve BatchPredictionJob from. If not set,project set in aiplatform.init will be used.

  • location – Optional[str] = None,Optional. location to retrieve BatchPredictionJob from. If not set,location set in aiplatform.init will be used.

  • credentials – Optional[auth_credentials.Credentials] = None,Custom credentials to use. If not set, credentials set inaiplatform.init will be used.

cancel()

Cancels this Job.

Success of cancellation is not guaranteed. Use Job.stateproperty to verify if cancellation was successful.

client_class()

alias ofgoogle.cloud.aiplatform.utils.JobClientWithOverride

property completion_stats(: Optional[[google.cloud.aiplatform_v1.types.completion_stats.CompletionStats](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.CompletionStats)_ )

Statistics on completed and failed prediction instances.

classmethod create(job_display_name:str, model_name:Union[str, google.cloud.aiplatform.models.Model], instances_format:str = 'jsonl', predictions_format:str = 'jsonl', gcs_source:Optional[Union[str,Sequence[str]]] = None, bigquery_source:Optional[str] = None, gcs_destination_prefix:Optional[str] = None, bigquery_destination_prefix:Optional[str] = None, model_parameters:Optional[Dict] = None, machine_type:Optional[str] = None, accelerator_type:Optional[str] = None, accelerator_count:Optional[int] = None, starting_replica_count:Optional[int] = None, max_replica_count:Optional[int] = None, generate_explanation:Optional[bool] = False, explanation_metadata:Optional[google.cloud.aiplatform_v1.types.explanation_metadata.ExplanationMetadata] = None, explanation_parameters:Optional[google.cloud.aiplatform_v1.types.explanation.ExplanationParameters] = None, labels:Optional[Dict[str,str]] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, encryption_spec_key_name:Optional[str] = None, sync:bool = True, create_request_timeout:Optional[float] = None, batch_size:Optional[int] = None, model_monitoring_objective_config:Optional[google.cloud.aiplatform.model_monitoring.objective.ObjectiveConfig] = None, model_monitoring_alert_config:Optional[google.cloud.aiplatform.model_monitoring.alert.AlertConfig] = None, analysis_instance_schema_uri:Optional[str] = None, service_account:Optional[str] = None)

Create a batch prediction job.

• Parameters

  • job_display_name (str) – Required. The user-defined name of the BatchPredictionJob.The name can be up to 128 characters long and can be consistof any UTF-8 characters.

  • model_name (Union[str, **aiplatform.Model]) – Required. A fully-qualified model resource name or model ID.Example: “projects/123/locations/us-central1/models/456” or“456” when project and location are initialized or passed.May optionally contain a version ID or alias in{model_name}@{version} form.

    Or an instance of aiplatform.Model.

  • instances_format (str) – Required. The format in which instances are provided. Must be oneof the formats listed in Model.supported_input_storage_formats.Default is “jsonl” when using gcs_source. If a bigquery_sourceis provided, this is overridden to “bigquery”.

  • predictions_format (str) – Required. The format in which Vertex AI outputs thepredictions, must be one of the formats specified inModel.supported_output_storage_formats.Default is “jsonl” when using gcs_destination_prefix. If abigquery_destination_prefix is provided, this is overridden to“bigquery”.

  • gcs_source (Optional[Sequence[str]]) – Google Cloud Storage URI(-s) to your instances to runbatch prediction on. They must match instances_format.

  • bigquery_source (Optional[str]) – BigQuery URI to a table, up to 2000 characters long. For example:bq://projectId.bqDatasetId.bqTableId

  • gcs_destination_prefix (Optional[str]) – The Google Cloud Storage location of the directory where theoutput is to be written to. In the given directory a newdirectory is created. Its name isprediction-<model-display-name>-<job-create-time>, wheretimestamp is in YYYY-MM-DDThh:mm:ss.sssZ ISO-8601 format.Inside of it filespredictions_0001.<extension>,predictions_0002.<extension>, …,predictions_N.<extension> are created where<extension> depends on chosenpredictions_format,and N may equal 0001 and depends on the total number ofsuccessfully predicted instances. If the Model has bothinstance andprediction schemata defined then each suchfile contains predictions as per thepredictions_format.If prediction for any instance failed (partially orcompletely), then an additionalerrors_0001.<extension>,errors_0002.<extension>,…,errors_N.<extension>files are created (N depends on total number of failedpredictions). These files contain the failed instances, asper their schema, followed by an additionalerror fieldwhich as value has\google.rpc.Status<Status>\__containing onlycode andmessage fields.

  • bigquery_destination_prefix (Optional[str]) – The BigQuery project or dataset location where the output isto be written to. If project is provided, a new dataset iscreated with nameprediction_<model-display-name>_<job-create-time> whereis made BigQuery-dataset-name compatible (for example, mostspecial characters become underscores), and timestamp is inYYYY_MM_DDThh_mm_ss_sssZ “based on ISO-8601” format. In thedataset two tables will be created,predictions, anderrors. If the Model has both[instance][google.cloud.aiplatform.v1.PredictSchemata.instance_schema_uri]and[prediction][google.cloud.aiplatform.v1.PredictSchemata.parameters_schema_uri]schemata defined then the tables have columns as follows:Thepredictions table contains instances for which theprediction succeeded, it has columns as per a concatenationof the Model’s instance and prediction schemata. Theerrors table contains rows for which the prediction hasfailed, it has instance columns, as per the instance schema,followed by a single “errors” column, which as values has[google.rpc.Status][google.rpc.Status] represented as aSTRUCT, and containing onlycode andmessage.

  • model_parameters (Optional[Dict]) – The parameters that govern the predictions. The schema ofthe parameters may be specified via the Model’s parameters_schema_uri.

  • machine_type (Optional[str]) – The type of machine for running batch prediction ondedicated resources. Not specifying machine type will result inbatch prediction job being run with automatic resources.

  • accelerator_type (Optional[str]) – The type of accelerator(s) that may be attachedto the machine as per accelerator_count. Only used ifmachine_type is set.

  • accelerator_count (Optional[int]) – The number of accelerators to attach to themachine_type. Only used if machine_type is set.

  • starting_replica_count (Optional[int]) – The number of machine replicas used at the start of the batchoperation. If not set, Vertex AI decides starting number, notgreater than max_replica_count. Only used if machine_type isset.

  • max_replica_count (Optional[int]) – The maximum number of machine replicas the batch operation maybe scaled to. Only used if machine_type is set.Default is 10.

  • generate_explanation (bool) – Optional. Generate explanation along with the batch predictionresults. This will cause the batch prediction output to includeexplanations based on the prediction_format:

    * bigquery: output includes a column named explanation. The valueis a struct that conforms to the [aiplatform.gapic.Explanation] object.

>     * jsonl: The JSON objects on each line include an additional entry>     keyed explanation. The value of the entry is a JSON object that>     conforms to the [aiplatform.gapic.Explanation] object.>     * csv: Generating explanations for CSV format is not supported.* **explanation_metadata** (*aiplatform.explain.ExplanationMetadata*) – Optional. Explanation metadata configuration for this BatchPredictionJob.Can be specified only if generate_explanation is set to True.This value overrides the value of Model.explanation_metadata.All fields of explanation_metadata are optional in the request. Ifa field of the explanation_metadata object is not populated, thecorresponding field of the Model.explanation_metadata object is inherited.For more details, see Ref docs <http://tinyurl.com/1igh60kt>* **explanation_parameters** (*aiplatform.explain.ExplanationParameters*) – Optional. Parameters to configure explaining for Model’s predictions.Can be specified only if generate_explanation is set to True.This value overrides the value of Model.explanation_parameters.All fields of explanation_parameters are optional in the request. Ifa field of the explanation_parameters object is not populated, thecorresponding field of the Model.explanation_parameters object is inherited.For more details, see Ref docs <http://tinyurl.com/1an4zake>* **labels** (*Dict**[*[*str*](https://docs.python.org/3/library/stdtypes.html#str)*, *[*str*](https://docs.python.org/3/library/stdtypes.html#str)*]*) – Optional. The labels with user-defined metadata to organize yourBatchPredictionJobs. Label keys and values can be no longer than64 characters (Unicode codepoints), can only contain lowercaseletters, numeric characters, underscores and dashes.International characters are allowed. See [https://goo.gl/xmQnxf](https://goo.gl/xmQnxf)for more information and examples of labels.* **credentials** (*Optional**[**auth_credentials.Credentials**]*) – Custom credentials to use to create this batch predictionjob. Overrides credentials set in aiplatform.init.* **encryption_spec_key_name** (*Optional**[*[*str*](https://docs.python.org/3/library/stdtypes.html#str)*]*) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the job. Has theform:`projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key`.The key needs to be in the same region as where the computeresource is created.If this is set, then allresources created by the BatchPredictionJob willbe encrypted with the provided encryption key.Overrides encryption_spec_key_name set in aiplatform.init.* **sync** ([*bool*](https://docs.python.org/3/library/functions.html#bool)) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.* **create_request_timeout** ([*float*](https://docs.python.org/3/library/functions.html#float)) – Optional. The timeout for the create request in seconds.* **batch_size** ([*int*](https://docs.python.org/3/library/functions.html#int)) – Optional. The number of the records (e.g. instances) of the operation given in each batchto a machine replica. Machine type, and size of a single record should be consideredwhen setting this parameter, higher value speeds up the batch operation’s execution,but too high value will result in a whole batch not fitting in a machine’s memory,and the whole operation will fail.The default value is 64.* **model_monitoring_objective_config** (*aiplatform.model_monitoring.ObjectiveConfig*) – Optional. The objective config for model monitoring. Passing this parameter enablesmonitoring on the model associated with this batch prediction job.* **model_monitoring_alert_config** (*aiplatform.model_monitoring.EmailAlertConfig*) – Optional. Configures how model monitoring alerts are sent to the user. Right nowonly email alert is supported.* **analysis_instance_schema_uri** ([*str*](https://docs.python.org/3/library/stdtypes.html#str)) – Optional. Only applicable if model_monitoring_objective_config is also passed.This parameter specifies the YAML schema file uri describing the format of a singleinstance that you want Tensorflow Data Validation (TFDV) toanalyze. If this field is empty, all the feature data types areinferred from predict_instance_schema_uri, meaning that TFDVwill use the data in the exact format as prediction request/response.If there are any data type differences between predict instanceand TFDV instance, this field can be used to override the schema.For models trained with Vertex AI, this field must be set as all thefields in predict instance formatted as string.* **service_account** ([*str*](https://docs.python.org/3/library/stdtypes.html#str)) – Optional. Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.

• Returns

  Instantiated representation of the created batch prediction job.

• Return type

  (jobs.BatchPredictionJob)

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Time when the Job resource entered the JOB_STATE_SUCCEEDED,JOB_STATE_FAILED, or JOB_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Detailed error info for this Job resource. Only populated when theJob’s state is JOB_STATE_FAILED or JOB_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

iter_outputs(bq_max_results:Optional[int] = 100)

Returns an Iterable object to traverse the output files, either alist of GCS Blobs or a BigQuery RowIterator depending on the outputconfig set when the BatchPredictionJob was created.

• Parameters

  bq_max_results – Optional[int] = 100 Limit on rows to retrieve from prediction table in BigQuery dataset. Only used when retrieving predictions from a bigquery_destination_prefix. Default is 100.

• Returns

  Either a list of GCS Blob objects within the prediction output directory or an iterable BigQuery RowIterator with predictions.

• Return type

  Union[Iterable[storage.Blob], Iterable[bigquery.table.RowIterator]]

• Raises

  • RuntimeError – If BatchPredictionJob is in a JobState other than SUCCEEDED, since outputs cannot be retrieved until the Job has finished.

  • NotImplementedError – If BatchPredictionJob succeeded and output_info does not have a GCS or BQ output provided.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

List all instances of this Job Resource.

Example Usage:

aiplatform.BatchPredictionJobs.list(

filter=’state=”JOB_STATE_SUCCEEDED” AND display_name=”my_job”’,

)

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names both snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields to order by, sorted inascending order. Use “desc” after a field name for descending.Supported fields: display_name, create_time, update_time

  • project (str) – Optional. Project to retrieve list from. If not set, projectset in aiplatform.init will be used.

  • location (str) – Optional. Location to retrieve list from. If not set, locationset in aiplatform.init will be used.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to retrieve list. Overridescredentials set in aiplatform.init.

• Returns

  List[VertexAiResourceNoun] - A list of Job resource objects.

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property output_info(: Optional[[google.cloud.aiplatform_v1.types.batch_prediction_job.BatchPredictionJob.OutputInfo](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.BatchPredictionJob.OutputInfo)_ )

Information describing the output of this job, including output locationinto which prediction output is written.

This is only available for batch prediction jobs that have run successfully.

property partial_failures(: Optional[Sequence[google.rpc.status_pb2.Status] )

Partial failures encountered. For example, single files that can’t be read.This field never exceeds 20 entries. Status details fields contain standardGCP error details.

property preview(: google.cloud.aiplatform.base.PreviewClas )

Exposes features available in preview for this class.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

property start_time(: Optional[datetime.datetime )

Time when the Job resource entered the JOB_STATE_RUNNING for thefirst time.

property state(: [google.cloud.aiplatform_v1.types.job_state.JobState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.JobState_ )

Fetch Job again and return the current JobState.

• Returns

  Enum that describes the state of a Vertex AI job.

• Return type

  state (job_state.JobState)

classmethod submit(*, job_display_name:Optional[str] = None, model_name:Union[str, google.cloud.aiplatform.models.Model], instances_format:str = 'jsonl', predictions_format:str = 'jsonl', gcs_source:Optional[Union[str,Sequence[str]]] = None, bigquery_source:Optional[str] = None, gcs_destination_prefix:Optional[str] = None, bigquery_destination_prefix:Optional[str] = None, model_parameters:Optional[Dict] = None, machine_type:Optional[str] = None, accelerator_type:Optional[str] = None, accelerator_count:Optional[int] = None, starting_replica_count:Optional[int] = None, max_replica_count:Optional[int] = None, generate_explanation:Optional[bool] = False, explanation_metadata:Optional[google.cloud.aiplatform_v1.types.explanation_metadata.ExplanationMetadata] = None, explanation_parameters:Optional[google.cloud.aiplatform_v1.types.explanation.ExplanationParameters] = None, labels:Optional[Dict[str,str]] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, encryption_spec_key_name:Optional[str] = None, create_request_timeout:Optional[float] = None, batch_size:Optional[int] = None, model_monitoring_objective_config:Optional[google.cloud.aiplatform.model_monitoring.objective.ObjectiveConfig] = None, model_monitoring_alert_config:Optional[google.cloud.aiplatform.model_monitoring.alert.AlertConfig] = None, analysis_instance_schema_uri:Optional[str] = None, service_account:Optional[str] = None)

Sumbit a batch prediction job (not waiting for completion).

• Parameters

  • job_display_name (str) – Required. The user-defined name of the BatchPredictionJob.The name can be up to 128 characters long and can be consistof any UTF-8 characters.

  • model_name (Union[str, **aiplatform.Model]) – Required. A fully-qualified model resource name or model ID.Example: “projects/123/locations/us-central1/models/456” or“456” when project and location are initialized or passed.May optionally contain a version ID or alias in{model_name}@{version} form.

    Or an instance of aiplatform.Model.

  • instances_format (str) – Required. The format in which instances are provided. Must be oneof the formats listed in Model.supported_input_storage_formats.Default is “jsonl” when using gcs_source. If a bigquery_sourceis provided, this is overridden to “bigquery”.

  • predictions_format (str) – Required. The format in which Vertex AI outputs thepredictions, must be one of the formats specified inModel.supported_output_storage_formats.Default is “jsonl” when using gcs_destination_prefix. If abigquery_destination_prefix is provided, this is overridden to“bigquery”.

  • gcs_source (Optional[Sequence[str]]) – Google Cloud Storage URI(-s) to your instances to runbatch prediction on. They must match instances_format.

  • bigquery_source (Optional[str]) – BigQuery URI to a table, up to 2000 characters long. For example:bq://projectId.bqDatasetId.bqTableId

  • gcs_destination_prefix (Optional[str]) – The Google Cloud Storage location of the directory where theoutput is to be written to. In the given directory a newdirectory is created. Its name isprediction-<model-display-name>-<job-create-time>, wheretimestamp is in YYYY-MM-DDThh:mm:ss.sssZ ISO-8601 format.Inside of it filespredictions_0001.<extension>,predictions_0002.<extension>, …,predictions_N.<extension> are created where<extension> depends on chosenpredictions_format,and N may equal 0001 and depends on the total number ofsuccessfully predicted instances. If the Model has bothinstance andprediction schemata defined then each suchfile contains predictions as per thepredictions_format.If prediction for any instance failed (partially orcompletely), then an additionalerrors_0001.<extension>,errors_0002.<extension>,…,errors_N.<extension>files are created (N depends on total number of failedpredictions). These files contain the failed instances, asper their schema, followed by an additionalerror fieldwhich as value has\google.rpc.Status<Status>\__containing onlycode andmessage fields.

  • bigquery_destination_prefix (Optional[str]) – The BigQuery project or dataset location where the output isto be written to. If project is provided, a new dataset iscreated with nameprediction_<model-display-name>_<job-create-time> whereis made BigQuery-dataset-name compatible (for example, mostspecial characters become underscores), and timestamp is inYYYY_MM_DDThh_mm_ss_sssZ “based on ISO-8601” format. In thedataset two tables will be created,predictions, anderrors. If the Model has both[instance][google.cloud.aiplatform.v1.PredictSchemata.instance_schema_uri]and[prediction][google.cloud.aiplatform.v1.PredictSchemata.parameters_schema_uri]schemata defined then the tables have columns as follows:Thepredictions table contains instances for which theprediction succeeded, it has columns as per a concatenationof the Model’s instance and prediction schemata. Theerrors table contains rows for which the prediction hasfailed, it has instance columns, as per the instance schema,followed by a single “errors” column, which as values has[google.rpc.Status][google.rpc.Status] represented as aSTRUCT, and containing onlycode andmessage.

  • model_parameters (Optional[Dict]) – The parameters that govern the predictions. The schema ofthe parameters may be specified via the Model’s parameters_schema_uri.

  • machine_type (Optional[str]) – The type of machine for running batch prediction ondedicated resources. Not specifying machine type will result inbatch prediction job being run with automatic resources.

  • accelerator_type (Optional[str]) – The type of accelerator(s) that may be attachedto the machine as per accelerator_count. Only used ifmachine_type is set.

  • accelerator_count (Optional[int]) – The number of accelerators to attach to themachine_type. Only used if machine_type is set.

  • starting_replica_count (Optional[int]) – The number of machine replicas used at the start of the batchoperation. If not set, Vertex AI decides starting number, notgreater than max_replica_count. Only used if machine_type isset.

  • max_replica_count (Optional[int]) – The maximum number of machine replicas the batch operation maybe scaled to. Only used if machine_type is set.Default is 10.

  • generate_explanation (bool) – Optional. Generate explanation along with the batch predictionresults. This will cause the batch prediction output to includeexplanations based on the prediction_format:

    * bigquery: output includes a column named explanation. The valueis a struct that conforms to the [aiplatform.gapic.Explanation] object.

>     * jsonl: The JSON objects on each line include an additional entry>     keyed explanation. The value of the entry is a JSON object that>     conforms to the [aiplatform.gapic.Explanation] object.>     * csv: Generating explanations for CSV format is not supported.* **explanation_metadata** (*aiplatform.explain.ExplanationMetadata*) – Optional. Explanation metadata configuration for this BatchPredictionJob.Can be specified only if generate_explanation is set to True.This value overrides the value of Model.explanation_metadata.All fields of explanation_metadata are optional in the request. Ifa field of the explanation_metadata object is not populated, thecorresponding field of the Model.explanation_metadata object is inherited.For more details, see Ref docs <http://tinyurl.com/1igh60kt>* **explanation_parameters** (*aiplatform.explain.ExplanationParameters*) – Optional. Parameters to configure explaining for Model’s predictions.Can be specified only if generate_explanation is set to True.This value overrides the value of Model.explanation_parameters.All fields of explanation_parameters are optional in the request. Ifa field of the explanation_parameters object is not populated, thecorresponding field of the Model.explanation_parameters object is inherited.For more details, see Ref docs <http://tinyurl.com/1an4zake>* **labels** (*Dict**[*[*str*](https://docs.python.org/3/library/stdtypes.html#str)*, *[*str*](https://docs.python.org/3/library/stdtypes.html#str)*]*) – Optional. The labels with user-defined metadata to organize yourBatchPredictionJobs. Label keys and values can be no longer than64 characters (Unicode codepoints), can only contain lowercaseletters, numeric characters, underscores and dashes.International characters are allowed. See [https://goo.gl/xmQnxf](https://goo.gl/xmQnxf)for more information and examples of labels.* **credentials** (*Optional**[**auth_credentials.Credentials**]*) – Custom credentials to use to create this batch predictionjob. Overrides credentials set in aiplatform.init.* **encryption_spec_key_name** (*Optional**[*[*str*](https://docs.python.org/3/library/stdtypes.html#str)*]*) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the job. Has theform:`projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key`.The key needs to be in the same region as where the computeresource is created.If this is set, then allresources created by the BatchPredictionJob willbe encrypted with the provided encryption key.Overrides encryption_spec_key_name set in aiplatform.init.* **create_request_timeout** ([*float*](https://docs.python.org/3/library/functions.html#float)) – Optional. The timeout for the create request in seconds.* **batch_size** ([*int*](https://docs.python.org/3/library/functions.html#int)) – Optional. The number of the records (e.g. instances) of the operation given in each batchto a machine replica. Machine type, and size of a single record should be consideredwhen setting this parameter, higher value speeds up the batch operation’s execution,but too high value will result in a whole batch not fitting in a machine’s memory,and the whole operation will fail.The default value is 64.* **model_monitoring_objective_config** (*aiplatform.model_monitoring.ObjectiveConfig*) – Optional. The objective config for model monitoring. Passing this parameter enablesmonitoring on the model associated with this batch prediction job.* **model_monitoring_alert_config** (*aiplatform.model_monitoring.EmailAlertConfig*) – Optional. Configures how model monitoring alerts are sent to the user. Right nowonly email alert is supported.* **analysis_instance_schema_uri** ([*str*](https://docs.python.org/3/library/stdtypes.html#str)) – Optional. Only applicable if model_monitoring_objective_config is also passed.This parameter specifies the YAML schema file uri describing the format of a singleinstance that you want Tensorflow Data Validation (TFDV) toanalyze. If this field is empty, all the feature data types areinferred from predict_instance_schema_uri, meaning that TFDVwill use the data in the exact format as prediction request/response.If there are any data type differences between predict instanceand TFDV instance, this field can be used to override the schema.For models trained with Vertex AI, this field must be set as all thefields in predict instance formatted as string.* **service_account** ([*str*](https://docs.python.org/3/library/stdtypes.html#str)) – Optional. Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.

• Returns

  Instantiated representation of the created batch prediction job.

• Return type

  (jobs.BatchPredictionJob)

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_completion()

Waits for job to complete.

• Raises

  RuntimeError – If job failed or cancelled.

wait_for_resource_creation()

Waits until resource has been created.

class google.cloud.aiplatform.CustomContainerTrainingJob(display_name:str, container_uri:str, command:Optional[Sequence[str]] = None, model_serving_container_image_uri:Optional[str] = None, model_serving_container_predict_route:Optional[str] = None, model_serving_container_health_route:Optional[str] = None, model_serving_container_command:Optional[Sequence[str]] = None, model_serving_container_args:Optional[Sequence[str]] = None, model_serving_container_environment_variables:Optional[Dict[str,str]] = None, model_serving_container_ports:Optional[Sequence[int]] = None, model_description:Optional[str] = None, model_instance_schema_uri:Optional[str] = None, model_parameters_schema_uri:Optional[str] = None, model_prediction_schema_uri:Optional[str] = None, explanation_metadata:Optional[google.cloud.aiplatform_v1.types.explanation_metadata.ExplanationMetadata] = None, explanation_parameters:Optional[google.cloud.aiplatform_v1.types.explanation.ExplanationParameters] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None, staging_bucket:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._CustomTrainingJob

Class to launch a Custom Training Job in Vertex AI using aContainer.

Constructs a Custom Container Training Job.

job = aiplatform.CustomContainerTrainingJob(

display_name=’test-train’,container_uri=’gcr.io/my_project_id/my_image_name:tag’,command=[‘python3’, ‘run_script.py’]model_serving_container_image_uri=’gcr.io/my-trainer/serving:1’,model_serving_container_predict_route=’predict’,model_serving_container_health_route=’metadata,labels={‘key’: ‘value’},

)

Usage with Dataset:

ds = aiplatform.TabularDataset(

‘projects/my-project/locations/us-central1/datasets/12345’)

job.run(

ds,replica_count=1,model_display_name=’my-trained-model’,model_labels={‘key’: ‘value’},

)

Usage without Dataset:

job.run(replica_count=1, model_display_name=’my-trained-model)

To ensure your model gets saved in Vertex AI, write your saved model toos.environ[“AIP_MODEL_DIR”] in your provided training script.

• Parameters

  • display_name (str) – Required. The user-defined name of this TrainingPipeline.

  • container_uri (str) – Required: Uri of the training container image in the GCR.

  • command (Sequence[str]) – The command to be invoked when the container is started.It overrides the entrypoint instruction in Dockerfile when provided

  • model_serving_container_image_uri (str) – If the training produces a managed Vertex AI Model, the URI of theModel serving container suitable for serving the model produced by thetraining script.

  • model_serving_container_predict_route (str) – If the training produces a managed Vertex AI Model, An HTTP path tosend prediction requests to the container, and which must be supportedby it. If not specified a default HTTP path will be used by Vertex AI.

  • model_serving_container_health_route (str) – If the training produces a managed Vertex AI Model, an HTTP path tosend health check requests to the container, and which must be supportedby it. If not specified a standard HTTP path will be used by AIPlatform.

  • model_serving_container_command (Sequence[str]) – The command with which the container is run. Not executed within ashell. The Docker image’s ENTRYPOINT is used if this is not provided.Variable references $(VAR_NAME) are expanded using the container’senvironment. If a variable cannot be resolved, the reference in theinput string will be unchanged. The $(VAR_NAME) syntax can be escapedwith a double $$, ie: $$(VAR_NAME). Escaped references will never beexpanded, regardless of whether the variable exists or not.

  • model_serving_container_args (Sequence[str]) – The arguments to the command. The Docker image’s CMD is used if this isnot provided. Variable references $(VAR_NAME) are expanded using thecontainer’s environment. If a variable cannot be resolved, the referencein the input string will be unchanged. The $(VAR_NAME) syntax can beescaped with a double $$, ie: $$(VAR_NAME). Escaped references willnever be expanded, regardless of whether the variable exists or not.

  • model_serving_container_environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – The environment variables that are to be present in the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.

  • model_serving_container_ports (Sequence[int]) – Declaration of ports that are exposed by the container. This field isprimarily informational, it gives Vertex AI information about thenetwork connections the container uses. Listing or not a port here hasno impact on whether the port is actually exposed, any port listening onthe default “0.0.0.0” address inside a container will be accessible fromthe network.

  • model_description (str) – The description of the Model.

  • model_instance_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the format of a single instance, whichare used inPredictRequest.instances,ExplainRequest.instancesandBatchPredictionJob.input_config.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform. Note: The URI given on output will be immutableand probably different, including the URI scheme, than theone given on input. The output URI will point to a locationwhere the user only has a read access.

  • model_parameters_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the parameters of prediction andexplanation viaPredictRequest.parameters,ExplainRequest.parametersandBatchPredictionJob.model_parameters.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform, if no parameters are supported it is set to anempty string. Note: The URI given on output will beimmutable and probably different, including the URI scheme,than the one given on input. The output URI will point to alocation where the user only has a read access.

  • model_prediction_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the format of a single predictionproduced by this Model, which are returned viaPredictResponse.predictions,ExplainResponse.explanations,andBatchPredictionJob.output_config.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform. Note: The URI given on output will be immutableand probably different, including the URI scheme, than theone given on input. The output URI will point to a locationwhere the user only has a read access.

  • explanation_metadata (explain.ExplanationMetadata) – Optional. Metadata describing the Model’s input and output forexplanation. explanation_metadata is optional whileexplanation_parameters must be specified when used.For more details, see Ref docshttp://tinyurl.com/1igh60kt

  • explanation_parameters (explain.ExplanationParameters) – Optional. Parameters to configure explaining for Model’spredictions.For more details, see Ref docshttp://tinyurl.com/1an4zake

  • project (str) – Project to run training in. Overrides project set in aiplatform.init.

  • location (str) – Location to run training in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Custom credentials to use to run call training service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize TrainingPipelines.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the training pipeline. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, this TrainingPipeline will be secured by this key.

    Note: Model trained by this TrainingPipeline is also securedby this key ifmodel_to_upload is not set separately.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, the trained Model will be secured by this key.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • staging_bucket (str) – Bucket used to stage source and training artifacts. Overridesstaging_bucket set in aiplatform.init.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property network(: Optional[str )

The full name of the Google Compute Enginenetwork to which thisCustomTrainingJob should be peered.

Specify the name of the network using the formatprojects/{project}/global/networks/{network}. Replace {project} withthe project number, such as 12345, and {network} with a network name.

Before specifying a network, private services access must be configuredfor the network. If private services access isn’t configured, thenthe custom training job can’t be peered with a network.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset:Optional[Union[google.cloud.aiplatform.datasets.image_dataset.ImageDataset, google.cloud.aiplatform.datasets.tabular_dataset.TabularDataset, google.cloud.aiplatform.datasets.text_dataset.TextDataset, google.cloud.aiplatform.datasets.video_dataset.VideoDataset]] = None, annotation_schema_uri:Optional[str] = None, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, base_output_dir:Optional[str] = None, service_account:Optional[str] = None, network:Optional[str] = None, bigquery_destination:Optional[str] = None, args:Optional[List[Union[str,float,int]]] = None, environment_variables:Optional[Dict[str,str]] = None, replica_count:int = 1, machine_type:str = 'n1-standard-4', accelerator_type:str = 'ACCELERATOR_TYPE_UNSPECIFIED', accelerator_count:int = 0, boot_disk_type:str = 'pd-ssd', boot_disk_size_gb:int = 100, reduction_server_replica_count:int = 0, reduction_server_machine_type:Optional[str] = None, reduction_server_container_uri:Optional[str] = None, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, training_filter_split:Optional[str] = None, validation_filter_split:Optional[str] = None, test_filter_split:Optional[str] = None, predefined_split_column_name:Optional[str] = None, timestamp_split_column_name:Optional[str] = None, timeout:Optional[int] = None, restart_job_on_worker_restart:bool = False, enable_web_access:bool = False, enable_dashboard_access:bool = False, tensorboard:Optional[str] = None, sync=True, create_request_timeout:Optional[float] = None, disable_retries:bool = False, persistent_resource_id:Optional[str] = None, tpu_topology:Optional[str] = None, scheduling_strategy:Optional[google.cloud.aiplatform_v1.types.custom_job.Scheduling.Strategy] = None, reservation_affinity_type:Optional[Literal['NO_RESERVATION', 'ANY_RESERVATION', 'SPECIFIC_RESERVATION']] = None, reservation_affinity_key:Optional[str] = None, reservation_affinity_values:Optional[List[str]] = None, max_wait_duration:Optional[int] = None, psc_interface_config:Optional[google.cloud.aiplatform_v1.types.service_networking.PscInterfaceConfig] = None)

Runs the custom training job.

Distributed Training Support:If replica count = 1 then one chief replica will be provisioned. Ifreplica_count > 1 the remainder will be provisioned as a worker replica pool.ie: replica_count = 10 will result in 1 chief and 9 workersAll replicas have same machine_type, accelerator_type, and accelerator_count

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:Any of `training_fraction_split`, `validation_fraction_split` and`test_fraction_split` may optionally be provided, they must sum to up to 1. Ifthe provided ones sum to less than 1, the remainder is assigned to sets asdecided by Vertex AI. If none of the fractions are set, by default roughly 80%of data will be used for training, 10% for validation, and 10% for test.Data filter splits:Assigns input data to training, validation, and test setsbased on the given filters, data pieces not matched by anyfilter are ignored. Currently only supported for Datasetscontaining DataItems.If any of the filters in this message are to match nothing, thenthey can be set as ‘-’ (the minus sign).If using filter splits, all of `training_filter_split`, `validation_filter_split` and`test_filter_split` must be provided.Supported only for unstructured Datasets.Predefined splits:Assigns input data to training, validation, and test sets based on the value of a provided key.If using predefined splits, `predefined_split_column_name` must be provided.Supported only for tabular Datasets.Timestamp splits:Assigns input data to training, validation, and test setsbased on a provided timestamps. The youngest data pieces areassigned to training set, next to validation set, and the oldestto the test set.Supported only for tabular Datasets.

• Parameters

  • dataset (Union[datasets.ImageDataset,datasets.TabularDataset,datasets.TextDataset,datasets.VideoDataset]) – Vertex AI to fit this training against. Custom training script shouldretrieve datasets through passed in environment variables uris:

    os.environ[“AIP_TRAINING_DATA_URI”]os.environ[“AIP_VALIDATION_DATA_URI”]os.environ[“AIP_TEST_DATA_URI”]

    Additionally the dataset format is passed in as:

    os.environ[“AIP_DATA_FORMAT”]

  • annotation_schema_uri (str) – Google Cloud Storage URI points to a YAML file describingannotation schema. The schema is defined as an OpenAPI 3.0.2Schema Object The schema filesthat can be used here are found ings://google-cloud-aiplatform/schema/dataset/annotation/,note that the chosen schema must be consistent withmetadataof the Dataset specified bydataset_id.

    Only Annotations that both match this schema and belong toDataItems not ignored by the split method are used inrespectively training, validation or test role, depending onthe role of the DataItem they are on.

    When used in conjunction withannotations_filter,the Annotations used for training are filtered by bothannotations_filterandannotation_schema_uri.

  • model_display_name (str) – If the script produces a managed Vertex AI Model. The display name ofthe Model. The name can be up to 128 characters long and can be consistof any UTF-8 characters.

    If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • base_output_dir (str) – GCS output directory of job. If not provided atimestamped directory in the staging directory will be used.

    Vertex AI sets the following environment variables when it runs your training code:

    • AIP_MODEL_DIR: a Cloud Storage URI of a directory intended for saving model artifacts, i.e. <base_output_dir>/model/

    • AIP_CHECKPOINT_DIR: a Cloud Storage URI of a directory intended for saving checkpoints, i.e. <base_output_dir>/checkpoints/

    • AIP_TENSORBOARD_LOG_DIR: a Cloud Storage URI of a directory intended for saving TensorBoard logs, i.e. <base_output_dir>/logs/

  • service_account (str) – Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.

  • network (str) – The full name of the Compute Engine network to which the jobshould be peered. For example, projects/12345/global/networks/myVPC.Private services access must already be configured for the network.If left unspecified, the network set in aiplatform.init will be used.Otherwise, the job is not peered with any network.

  • bigquery_destination (str) – Provide this field if dataset is a BigQuery dataset.The BigQuery project location where the training data is tobe written to. In the given project a new dataset is createdwith namedataset_<dataset-id>_<annotation-type>_<timestamp-of-training-call>where timestamp is in YYYY_MM_DDThh_mm_ss_sssZ format. Alltraining input data will be written into that dataset. Inthe dataset three tables will be created,training,validation andtest.

    • AIP_DATA_FORMAT = “bigquery”.

    • AIP_TRAINING_DATA_URI =”bigquery_destination.dataset_*.training”

    • AIP_VALIDATION_DATA_URI = “bigquery_destination.dataset_*.validation”

    • AIP_TEST_DATA_URI = “bigquery_destination.dataset_*.test”

  • args (List[Unions[str, *[int](https://docs.python.org/3/library/functions.html#int),[float](https://docs.python.org/3/library/functions.html#float)]*]) – Command line arguments to be passed to the Python script.

  • environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Environment variables to be passed to the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.At most 10 environment variables can be specified.The Name of the environment variable must be unique.

    environment_variables = {

    ‘MY_KEY’: ‘MY_VALUE’

    }

  • replica_count (int) – The number of worker replicas. If replica count = 1 then one chiefreplica will be provisioned. If replica_count > 1 the remainder will beprovisioned as a worker replica pool.

  • machine_type (str) – The type of machine to use for training.

  • accelerator_type (str) – Hardware accelerator type. One of ACCELERATOR_TYPE_UNSPECIFIED,NVIDIA_TESLA_K80, NVIDIA_TESLA_P100, NVIDIA_TESLA_V100, NVIDIA_TESLA_P4,NVIDIA_TESLA_T4

  • accelerator_count (int) – The number of accelerators to attach to a worker replica.

  • boot_disk_type (str) – Type of the boot disk, default is pd-ssd.Valid values: pd-ssd (Persistent Disk Solid State Drive) orpd-standard (Persistent Disk Hard Disk Drive).

  • boot_disk_size_gb (int) – Size in GB of the boot disk, default is 100GB.boot disk size must be within the range of [100, 64000].

  • reduction_server_replica_count (int) – The number of reduction server replicas, default is 0.

  • reduction_server_machine_type (str) – Optional. The type of machine to use for reduction server.

  • reduction_server_container_uri (str) – Optional. The Uri of the reduction server container image.See details:https://cloud.google.com/vertex-ai/docs/training/distributed-training#reduce_training_time_with_reduction_server

  • training_fraction_split (float) – Optional. The fraction of the input data that is to be used to trainthe Model. This is ignored if Dataset is not provided.

  • validation_fraction_split (float) – Optional. The fraction of the input data that is to be used to validatethe Model. This is ignored if Dataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input data that is to be used to evaluatethe Model. This is ignored if Dataset is not provided.

  • training_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to train the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • validation_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to validate the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • test_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to test the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • predefined_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key (either the label’s value orvalue in the column) must be one of {training,validation,test}, and it defines to which set thegiven piece of data is assigned. If for a piece of data thekey is not present or has an invalid value, that piece isignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timestamp_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key values of the key (the values inthe column) must be in RFC 3339 date-time format, wheretime-offset = “Z” (e.g. 1985-04-12T23:20:50.52Z). If for apiece of data the key is not present or has an invalid value,that piece is ignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timeout (int) – The maximum job running time in seconds. The default is 7 days.

  • restart_job_on_worker_restart (bool) – Restarts the entire CustomJob if a workergets restarted. This feature can be used bydistributed training jobs that are not resilientto workers leaving and joining a job.

  • enable_web_access (bool) – Whether you want Vertex AI to enable interactive shell accessto training containers.https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell

  • enable_dashboard_access (bool) – Whether you want Vertex AI to enable access to the customized dashboardto training containers.

  • tensorboard (str) – Optional. The name of a Vertex AI[Tensorboard][google.cloud.aiplatform.v1beta1.Tensorboard]resource to which this CustomJob will upload Tensorboardlogs. Format:projects/{project}/locations/{location}/tensorboards/{tensorboard}

    The training script should write Tensorboard to following Vertex AI environmentvariable:

    AIP_TENSORBOARD_LOG_DIR

    service_account is required with provided tensorboard.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

  • disable_retries (bool) – Indicates if the job should retry for internal errors after thejob starts running. If True, overridesrestart_job_on_worker_restart to False.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network, CMEK, and node poolconfigs on the job should be consistent with those on thePersistentResource, otherwise, the job will be rejected.

  • tpu_topology (str) – Optional. Specifies the tpu topology to be used forTPU training job. This field is required for TPU v5 versions. Fordetails on the TPU topology, refer tohttps://cloud.google.com/tpu/docs/v5e#tpu-v5e-config. The topologymust be a supported value for the TPU machine type.

  • scheduling_strategy (gca_custom_job_compat.Scheduling.Strategy) – Optional. Indicates the job scheduling strategy.

  • reservation_affinity_type (str) – Optional. The type of reservation affinity. One of:* “NO_RESERVATION” : No reservation is used.* “ANY_RESERVATION” : Any reservation that matches machine speccan be used.* “SPECIFIC_RESERVATION” : A specific reservation must be useused. See reservation_affinity_key andreservation_affinity_values for how to specify the reservation.

  • reservation_affinity_key (str) – Optional. Corresponds to the label key of a reservation resource.To target a SPECIFIC_RESERVATION by name, usecompute.googleapis.com/reservation-name as the keyand specify the name of your reservation as its value.

  • reservation_affinity_values (List[str]) – Optional. Corresponds to the label values of a reservation resource.This must be the full resource name of the reservation.Format: ‘projects/{project_id_or_number}/zones/{zone}/reservations/{reservation_name}’

  • max_wait_duration (int) – This is the maximum duration that a job will wait for therequested resources to be provisioned in seconds. If set to 0,the job will wait indefinitely. The default is 30 minutes.

  • psc_interface_config (gca_service_networking.PscInterfaceConfig) – Optional. Configuration for Private Service Connect interfaceused for training.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run, staging_bucket has not been set, or model_display_name was provided but required arguments were not provided in constructor.

property start_time(: Optional[datetime.datetime )

Optional. The time when the training job first entered thePIPELINE_STATE_RUNNING state.

property state(: Optional[[google.cloud.aiplatform_v1.types.pipeline_state.PipelineState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.PipelineState)_ )

Current training state.

submit(dataset:Optional[Union[google.cloud.aiplatform.datasets.image_dataset.ImageDataset, google.cloud.aiplatform.datasets.tabular_dataset.TabularDataset, google.cloud.aiplatform.datasets.text_dataset.TextDataset, google.cloud.aiplatform.datasets.video_dataset.VideoDataset]] = None, annotation_schema_uri:Optional[str] = None, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, base_output_dir:Optional[str] = None, service_account:Optional[str] = None, network:Optional[str] = None, bigquery_destination:Optional[str] = None, args:Optional[List[Union[str,float,int]]] = None, environment_variables:Optional[Dict[str,str]] = None, replica_count:int = 1, machine_type:str = 'n1-standard-4', accelerator_type:str = 'ACCELERATOR_TYPE_UNSPECIFIED', accelerator_count:int = 0, boot_disk_type:str = 'pd-ssd', boot_disk_size_gb:int = 100, reduction_server_replica_count:int = 0, reduction_server_machine_type:Optional[str] = None, reduction_server_container_uri:Optional[str] = None, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, training_filter_split:Optional[str] = None, validation_filter_split:Optional[str] = None, test_filter_split:Optional[str] = None, predefined_split_column_name:Optional[str] = None, timestamp_split_column_name:Optional[str] = None, timeout:Optional[int] = None, restart_job_on_worker_restart:bool = False, enable_web_access:bool = False, enable_dashboard_access:bool = False, tensorboard:Optional[str] = None, sync=True, create_request_timeout:Optional[float] = None, disable_retries:bool = False, persistent_resource_id:Optional[str] = None, tpu_topology:Optional[str] = None, scheduling_strategy:Optional[google.cloud.aiplatform_v1.types.custom_job.Scheduling.Strategy] = None, reservation_affinity_type:Optional[Literal['NO_RESERVATION', 'ANY_RESERVATION', 'SPECIFIC_RESERVATION']] = None, reservation_affinity_key:Optional[str] = None, reservation_affinity_values:Optional[List[str]] = None, max_wait_duration:Optional[int] = None, psc_interface_config:Optional[google.cloud.aiplatform_v1.types.service_networking.PscInterfaceConfig] = None)

Submits the custom training job without blocking until completion.

Distributed Training Support:If replica count = 1 then one chief replica will be provisioned. Ifreplica_count > 1 the remainder will be provisioned as a worker replica pool.ie: replica_count = 10 will result in 1 chief and 9 workersAll replicas have same machine_type, accelerator_type, and accelerator_count

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:Any of `training_fraction_split`, `validation_fraction_split` and`test_fraction_split` may optionally be provided, they must sum to up to 1. Ifthe provided ones sum to less than 1, the remainder is assigned to sets asdecided by Vertex AI. If none of the fractions are set, by default roughly 80%of data will be used for training, 10% for validation, and 10% for test.Data filter splits:Assigns input data to training, validation, and test setsbased on the given filters, data pieces not matched by anyfilter are ignored. Currently only supported for Datasetscontaining DataItems.If any of the filters in this message are to match nothing, thenthey can be set as ‘-’ (the minus sign).If using filter splits, all of `training_filter_split`, `validation_filter_split` and`test_filter_split` must be provided.Supported only for unstructured Datasets.Predefined splits:Assigns input data to training, validation, and test sets based on the value of a provided key.If using predefined splits, `predefined_split_column_name` must be provided.Supported only for tabular Datasets.Timestamp splits:Assigns input data to training, validation, and test setsbased on a provided timestamps. The youngest data pieces areassigned to training set, next to validation set, and the oldestto the test set.Supported only for tabular Datasets.

• Parameters

  • dataset (Union[datasets.ImageDataset,datasets.TabularDataset,datasets.TextDataset,datasets.VideoDataset]) – Vertex AI to fit this training against. Custom training script shouldretrieve datasets through passed in environment variables uris:

    os.environ[“AIP_TRAINING_DATA_URI”]os.environ[“AIP_VALIDATION_DATA_URI”]os.environ[“AIP_TEST_DATA_URI”]

    Additionally the dataset format is passed in as:

    os.environ[“AIP_DATA_FORMAT”]

  • annotation_schema_uri (str) – Google Cloud Storage URI points to a YAML file describingannotation schema. The schema is defined as an OpenAPI 3.0.2Schema Object The schema filesthat can be used here are found ings://google-cloud-aiplatform/schema/dataset/annotation/,note that the chosen schema must be consistent withmetadataof the Dataset specified bydataset_id.

    Only Annotations that both match this schema and belong toDataItems not ignored by the split method are used inrespectively training, validation or test role, depending onthe role of the DataItem they are on.

    When used in conjunction withannotations_filter,the Annotations used for training are filtered by bothannotations_filterandannotation_schema_uri.

  • model_display_name (str) – If the script produces a managed Vertex AI Model. The display name ofthe Model. The name can be up to 128 characters long and can be consistof any UTF-8 characters.

    If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • base_output_dir (str) – GCS output directory of job. If not provided atimestamped directory in the staging directory will be used.

    Vertex AI sets the following environment variables when it runs your training code:

    • AIP_MODEL_DIR: a Cloud Storage URI of a directory intended for saving model artifacts, i.e. <base_output_dir>/model/

    • AIP_CHECKPOINT_DIR: a Cloud Storage URI of a directory intended for saving checkpoints, i.e. <base_output_dir>/checkpoints/

    • AIP_TENSORBOARD_LOG_DIR: a Cloud Storage URI of a directory intended for saving TensorBoard logs, i.e. <base_output_dir>/logs/

  • service_account (str) – Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.

  • network (str) – The full name of the Compute Engine network to which the jobshould be peered. For example, projects/12345/global/networks/myVPC.Private services access must already be configured for the network.If left unspecified, the network set in aiplatform.init will be used.Otherwise, the job is not peered with any network.

  • bigquery_destination (str) – Provide this field if dataset is a BigQuery dataset.The BigQuery project location where the training data is tobe written to. In the given project a new dataset is createdwith namedataset_<dataset-id>_<annotation-type>_<timestamp-of-training-call>where timestamp is in YYYY_MM_DDThh_mm_ss_sssZ format. Alltraining input data will be written into that dataset. Inthe dataset three tables will be created,training,validation andtest.

    • AIP_DATA_FORMAT = “bigquery”.

    • AIP_TRAINING_DATA_URI =”bigquery_destination.dataset_*.training”

    • AIP_VALIDATION_DATA_URI = “bigquery_destination.dataset_*.validation”

    • AIP_TEST_DATA_URI = “bigquery_destination.dataset_*.test”

  • args (List[Unions[str, *[int](https://docs.python.org/3/library/functions.html#int),[float](https://docs.python.org/3/library/functions.html#float)]*]) – Command line arguments to be passed to the Python script.

  • environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Environment variables to be passed to the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.At most 10 environment variables can be specified.The Name of the environment variable must be unique.

    environment_variables = {

    ‘MY_KEY’: ‘MY_VALUE’

    }

  • replica_count (int) – The number of worker replicas. If replica count = 1 then one chiefreplica will be provisioned. If replica_count > 1 the remainder will beprovisioned as a worker replica pool.

  • machine_type (str) – The type of machine to use for training.

  • accelerator_type (str) – Hardware accelerator type. One of ACCELERATOR_TYPE_UNSPECIFIED,NVIDIA_TESLA_K80, NVIDIA_TESLA_P100, NVIDIA_TESLA_V100, NVIDIA_TESLA_P4,NVIDIA_TESLA_T4

  • accelerator_count (int) – The number of accelerators to attach to a worker replica.

  • boot_disk_type (str) – Type of the boot disk, default is pd-ssd.Valid values: pd-ssd (Persistent Disk Solid State Drive) orpd-standard (Persistent Disk Hard Disk Drive).

  • boot_disk_size_gb (int) – Size in GB of the boot disk, default is 100GB.boot disk size must be within the range of [100, 64000].

  • reduction_server_replica_count (int) – The number of reduction server replicas, default is 0.

  • reduction_server_machine_type (str) – Optional. The type of machine to use for reduction server.

  • reduction_server_container_uri (str) – Optional. The Uri of the reduction server container image.See details:https://cloud.google.com/vertex-ai/docs/training/distributed-training#reduce_training_time_with_reduction_server

  • training_fraction_split (float) – Optional. The fraction of the input data that is to be used to trainthe Model. This is ignored if Dataset is not provided.

  • validation_fraction_split (float) – Optional. The fraction of the input data that is to be used to validatethe Model. This is ignored if Dataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input data that is to be used to evaluatethe Model. This is ignored if Dataset is not provided.

  • training_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to train the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • validation_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to validate the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • test_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to test the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • predefined_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key (either the label’s value orvalue in the column) must be one of {training,validation,test}, and it defines to which set thegiven piece of data is assigned. If for a piece of data thekey is not present or has an invalid value, that piece isignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timestamp_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key values of the key (the values inthe column) must be in RFC 3339 date-time format, wheretime-offset = “Z” (e.g. 1985-04-12T23:20:50.52Z). If for apiece of data the key is not present or has an invalid value,that piece is ignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timeout (int) – The maximum job running time in seconds. The default is 7 days.

  • restart_job_on_worker_restart (bool) – Restarts the entire CustomJob if a workergets restarted. This feature can be used bydistributed training jobs that are not resilientto workers leaving and joining a job.

  • enable_web_access (bool) – Whether you want Vertex AI to enable interactive shell accessto training containers.https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell

  • enable_dashboard_access (bool) – Whether you want Vertex AI to enable access to the customized dashboardto training containers.

  • tensorboard (str) – Optional. The name of a Vertex AI[Tensorboard][google.cloud.aiplatform.v1beta1.Tensorboard]resource to which this CustomJob will upload Tensorboardlogs. Format:projects/{project}/locations/{location}/tensorboards/{tensorboard}

    The training script should write Tensorboard to following Vertex AI environmentvariable:

    AIP_TENSORBOARD_LOG_DIR

    service_account is required with provided tensorboard.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

  • disable_retries (bool) – Indicates if the job should retry for internal errors after thejob starts running. If True, overridesrestart_job_on_worker_restart to False.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network, CMEK, and node poolconfigs on the job should be consistent with those on thePersistentResource, otherwise, the job will be rejected.

  • tpu_topology (str) – Optional. Specifies the tpu topology to be used forTPU training job. This field is required for TPU v5 versions. Fordetails on the TPU topology, refer tohttps://cloud.google.com/tpu/docs/v5e#tpu-v5e-config. The topologymust be a supported value for the TPU machine type.

  • scheduling_strategy (gca_custom_job_compat.Scheduling.Strategy) – Optional. Indicates the job scheduling strategy.

  • reservation_affinity_type (str) – Optional. The type of reservation affinity. One of:* “NO_RESERVATION” : No reservation is used.* “ANY_RESERVATION” : Any reservation that matches machine speccan be used.* “SPECIFIC_RESERVATION” : A specific reservation must be useused. See reservation_affinity_key andreservation_affinity_values for how to specify the reservation.

  • reservation_affinity_key (str) – Optional. Corresponds to the label key of a reservation resource.To target a SPECIFIC_RESERVATION by name, usecompute.googleapis.com/reservation-name as the keyand specify the name of your reservation as its value.

  • reservation_affinity_values (List[str]) – Optional. Corresponds to the label values of a reservation resource.This must be the full resource name of the reservation.Format: ‘projects/{project_id_or_number}/zones/{zone}/reservations/{reservation_name}’

  • max_wait_duration (int) – This is the maximum duration that a job will wait for therequested resources to be provisioned in seconds. If set to 0,the job will wait indefinitely. The default is 30 minutes.

  • psc_interface_config (gca_service_networking.PscInterfaceConfig) – Optional. Configuration for Private Service Connect interfaceused for training.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model

• Raises

  RuntimeError – If Training job has already been run, staging_bucket has not been set, or model_display_name was provided but required arguments were not provided in constructor.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_resource_creation()

Waits until the resource has been created.

property web_access_uris(: Dict[str,str )

Returns the URIs used to access the custom training job.

class google.cloud.aiplatform.CustomJob(display_name:str, worker_pool_specs:Union[List[Dict],List[google.cloud.aiplatform_v1.types.custom_job.WorkerPoolSpec]], base_output_dir:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, encryption_spec_key_name:Optional[str] = None, staging_bucket:Optional[str] = None, persistent_resource_id:Optional[str] = None)

Bases:google.cloud.aiplatform.jobs._RunnableJob,google.cloud.aiplatform.base.PreviewMixin

Vertex AI Custom Job.

Constructs a Custom Job with Worker Pool Specs.

``

`Example usage:worker_pool_specs = [

{

“machine_spec”: {    “machine_type”: “n1-standard-4”,    “accelerator_type”: “NVIDIA_TESLA_K80”,    “accelerator_count”: 1,},“replica_count”: 1,“container_spec”: {> “image_uri”: container_image_uri,> “command”: [],> “args”: [],},

}

]

my_job = aiplatform.CustomJob(

display_name=’my_job’,worker_pool_specs=worker_pool_specs,labels={‘my_key’: ‘my_value’},

)

my_job.run()

``

`

For more information on configuring worker pool specs please visit:https://cloud.google.com/ai-platform-unified/docs/training/create-custom-job

• Parameters

  • display_name (str) – Required. The user-defined name of the HyperparameterTuningJob.The name can be up to 128 characters long and can be consistof any UTF-8 characters.

  • worker_pool_specs (Union[List[Dict], **List[aiplatform.gapic.WorkerPoolSpec]]) – Required. The spec of the worker pools including machine type and Docker image.Can provided as a list of dictionaries or list of WorkerPoolSpec proto messages.

  • base_output_dir (str) – Optional. GCS output directory of job. If not provided atimestamped directory in the staging directory will be used.

  • project (str) – Optional.Project to run the custom job in. Overrides project set in aiplatform.init.

  • location (str) – Optional.Location to run the custom job in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional.Custom credentials to use to run call custom job service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize CustomJobs.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • encryption_spec_key_name (str) – Optional.Customer-managed encryption key name for aCustomJob. If this is set, then all resourcescreated by the CustomJob will be encrypted withthe provided encryption key.

  • staging_bucket (str) – Optional. Bucket for produced custom job artifacts. Overridesstaging_bucket set in aiplatform.init.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network and CMEK configs onthe job should be consistent with those on the PersistentResource,otherwise, the job will be rejected.

• Raises

  RuntimeError – If staging bucket was not set using aiplatform.init and a staging bucket was not passed in.

cancel()

Cancels this Job.

Success of cancellation is not guaranteed. Use Job.stateproperty to verify if cancellation was successful.

client_class()

alias ofgoogle.cloud.aiplatform.utils.JobClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Time when the Job resource entered the JOB_STATE_SUCCEEDED,JOB_STATE_FAILED, or JOB_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Detailed error info for this Job resource. Only populated when theJob’s state is JOB_STATE_FAILED or JOB_STATE_CANCELLED.

classmethod from_local_script(display_name:str, script_path:str, container_uri:str, enable_autolog:bool = False, args:Optional[Sequence[str]] = None, requirements:Optional[Sequence[str]] = None, environment_variables:Optional[Dict[str,str]] = None, replica_count:int = 1, machine_type:str = 'n1-standard-4', accelerator_type:str = 'ACCELERATOR_TYPE_UNSPECIFIED', accelerator_count:int = 0, boot_disk_type:str = 'pd-ssd', boot_disk_size_gb:int = 100, reduction_server_replica_count:int = 0, reduction_server_machine_type:Optional[str] = None, reduction_server_container_uri:Optional[str] = None, base_output_dir:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, encryption_spec_key_name:Optional[str] = None, staging_bucket:Optional[str] = None, persistent_resource_id:Optional[str] = None, tpu_topology:Optional[str] = None)

Configures a custom job from a local script.

Example usage:

``

`job = aiplatform.CustomJob.from_local_script(

display_name=”my-custom-job”,script_path=”training_script.py”,container_uri=”gcr.io/cloud-aiplatform/training/tf-cpu.2-2:latest”,requirements=[“gcsfs==0.7.1”],replica_count=1,args=[’–dataset’, ‘gs://my-bucket/my-dataset’,‘–model_output_uri’, ‘gs://my-bucket/model’]labels={‘my_key’: ‘my_value’},

)

job.run()

``

`

• Parameters

  • display_name (str) – Required. The user-defined name of this CustomJob.

  • script_path (str) – Required. Local path to training script.

  • container_uri (str) – Required. Uri of the training container image to use for custom job.Support images in Artifact Registry, Container Registry, or Docker Hub.Vertex AI provides a wide range of executor images with pre-installedpackages to meet users’ various use cases. See the list of pre-built containersfor traininghttps://cloud.google.com/vertex-ai/docs/training/pre-built-containers.If not using image from this list, please make sure python3 and pip3 are installed in your container.

  • enable_autolog (bool) – Optional. If True, the Vertex Experiments autologging feature will beenabled in the CustomJob. Note that this will wrap your training scriptwith some autologging-related code.

  • args (Optional[Sequence[str]]) – Optional. Command line arguments to be passed to the Python task.

  • requirements (Sequence[str]) – Optional. List of python packages dependencies of script.

  • environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. Environment variables to be passed to the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.At most 10 environment variables can be specified.The Name of the environment variable must be unique.

    environment_variables = {

    ‘MY_KEY’: ‘MY_VALUE’

    }

  • replica_count (int) – Optional. The number of worker replicas. If replica count = 1 then one chiefreplica will be provisioned. If replica_count > 1 the remainder will beprovisioned as a worker replica pool.

  • machine_type (str) – Optional. The type of machine to use for training.

  • accelerator_type (str) – Optional. Hardware accelerator type. One of ACCELERATOR_TYPE_UNSPECIFIED,NVIDIA_TESLA_K80, NVIDIA_TESLA_P100, NVIDIA_TESLA_V100, NVIDIA_TESLA_P4,NVIDIA_TESLA_T4

  • accelerator_count (int) – Optional. The number of accelerators to attach to a worker replica.

  • boot_disk_type (str) – Optional. Type of the boot disk, default is pd-ssd.Valid values: pd-ssd (Persistent Disk Solid State Drive) orpd-standard (Persistent Disk Hard Disk Drive).

  • boot_disk_size_gb (int) – Optional. Size in GB of the boot disk, default is 100GB.boot disk size must be within the range of [100, 64000].

  • reduction_server_replica_count (int) – The number of reduction server replicas, default is 0.

  • reduction_server_machine_type (str) – Optional. The type of machine to use for reduction server.

  • reduction_server_container_uri (str) – Optional. The Uri of the reduction server container image.See details:https://cloud.google.com/vertex-ai/docs/training/distributed-training#reduce_training_time_with_reduction_server

  • base_output_dir (str) – Optional. GCS output directory of job. If not provided atimestamped directory in the staging directory will be used.

  • project (str) – Optional. Project to run the custom job in. Overrides project set in aiplatform.init.

  • location (str) – Optional. Location to run the custom job in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to run call custom job service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize CustomJobs.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • encryption_spec_key_name (str) – Optional. Customer-managed encryption key name for aCustomJob. If this is set, then all resourcescreated by the CustomJob will be encrypted withthe provided encryption key.

  • staging_bucket (str) – Optional. Bucket for produced custom job artifacts. Overridesstaging_bucket set in aiplatform.init.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network, CMEK, and node poolconfigs on the job should be consistent with those on thePersistentResource, otherwise, the job will be rejected.

  • tpu_topology (str) – Optional. Specifies the tpu topology to be used forTPU training job. This field is required for TPU v5 versions. Fordetails on the TPU topology, refer tohttps://cloud.google.com/tpu/docs/v5e#tpu-v5e-config. The topologymust be a supported value for the TPU machine type.

• Raises

  RuntimeError – If staging bucket was not set using aiplatform.init and a staging bucket was not passed in.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Get a Vertex AI Job for the given resource_name.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. project to retrieve dataset from. If not set, projectset in aiplatform.init will be used.

  • location (str) – Optional. location to retrieve dataset from. If not set, locationset in aiplatform.init will be used.

  • credentials (auth_credentials.Credentials) – Custom credentials to use to upload this model. Overridescredentials set in aiplatform.init.

• Returns

  A Vertex AI Job.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

List all instances of this Job Resource.

Example Usage:

aiplatform.BatchPredictionJobs.list(

filter=’state=”JOB_STATE_SUCCEEDED” AND display_name=”my_job”’,

)

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names both snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields to order by, sorted inascending order. Use “desc” after a field name for descending.Supported fields: display_name, create_time, update_time

  • project (str) – Optional. Project to retrieve list from. If not set, projectset in aiplatform.init will be used.

  • location (str) – Optional. Location to retrieve list from. If not set, locationset in aiplatform.init will be used.

  • credentials (auth_credentials.Credentials) – Optional. Custom credentials to use to retrieve list. Overridescredentials set in aiplatform.init.

• Returns

  List[VertexAiResourceNoun] - A list of Job resource objects.

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property network(: Optional[str )

The full name of the Google Compute Enginenetwork to which thisCustomJob should be peered.

Takes the format projects/{project}/global/networks/{network}. Where{project} is a project number, as in 12345, and {network} is a network name.

Private services access must already be configured for the network. If leftunspecified, the CustomJob is not peered with any network.

property preview(: google.cloud.aiplatform.base.PreviewClas )

Exposes features available in preview for this class.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(service_account:Optional[str] = None, network:Optional[str] = None, timeout:Optional[int] = None, restart_job_on_worker_restart:bool = False, enable_web_access:bool = False, experiment:Optional[Union[google.cloud.aiplatform.metadata.experiment_resources.Experiment,str]] = None, experiment_run:Optional[Union[google.cloud.aiplatform.metadata.experiment_run_resource.ExperimentRun,str]] = None, tensorboard:Optional[str] = None, sync:bool = True, create_request_timeout:Optional[float] = None, disable_retries:bool = False, persistent_resource_id:Optional[str] = None, scheduling_strategy:Optional[google.cloud.aiplatform_v1.types.custom_job.Scheduling.Strategy] = None, max_wait_duration:Optional[int] = None, psc_interface_config:Optional[google.cloud.aiplatform_v1.types.service_networking.PscInterfaceConfig] = None)

Run this configured CustomJob.

• Parameters

  • service_account (str) – Optional. Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.

  • network (str) – Optional. The full name of the Compute Engine network to which the jobshould be peered. For example, projects/12345/global/networks/myVPC.Private services access must already be configured for the network.If left unspecified, the network set in aiplatform.init will be used.Otherwise, the job is not peered with any network.

  • timeout (int) – The maximum job running time in seconds. The default is 7 days.

  • restart_job_on_worker_restart (bool) – Restarts the entire CustomJob if a workergets restarted. This feature can be used bydistributed training jobs that are not resilientto workers leaving and joining a job.

  • enable_web_access (bool) – Whether you want Vertex AI to enable interactive shell accessto training containers.https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell

  • experiment (Union[aiplatform.Experiment, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The instance or name of an Experiment resource to whichthis CustomJob will upload training parameters and metrics.

    service_account is required with provided experiment.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • experiment_run (Union[aiplatform.ExperimentRun, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The instance or name of an ExperimentRun resource to whichthis CustomJob will upload training parameters and metrics.This arg can only be set when experiment is set. If ‘experiment’is set but ‘experiment_run` is not, an ExperimentRun resourcewill still be auto-generated.

  • tensorboard (str) – Optional. The name of a Vertex AI[Tensorboard][google.cloud.aiplatform.v1beta1.Tensorboard]resource to which this CustomJob will upload Tensorboardlogs. Format:projects/{project}/locations/{location}/tensorboards/{tensorboard}

    The training script should write Tensorboard to following Vertex AI environmentvariable:

    AIP_TENSORBOARD_LOG_DIR

    service_account is required with provided tensorboard.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill unblock and it will be executed in a concurrent Future.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

  • disable_retries (bool) – Indicates if the job should retry for internal errors after thejob starts running. If True, overridesrestart_job_on_worker_restart to False.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network, CMEK, and node poolconfigs on the job should be consistent with those on thePersistentResource, otherwise, the job will be rejected.

  • scheduling_strategy (gca_custom_job_compat.Scheduling.Strategy) – Optional. Indicates the job scheduling strategy.

  • max_wait_duration (int) – This is the maximum duration that a job will wait for therequested resources to be provisioned in seconds. If set to 0,the job will wait indefinitely. The default is 1 day.

  • psc_interface_config (gca_service_networking.PscInterfaceConfig) – Optional. Configuration for Private Service Connect interfaceused for training.

property start_time(: Optional[datetime.datetime )

Time when the Job resource entered the JOB_STATE_RUNNING for thefirst time.

property state(: [google.cloud.aiplatform_v1.types.job_state.JobState](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.JobState_ )

Fetch Job again and return the current JobState.

• Returns

  Enum that describes the state of a Vertex AI job.

• Return type

  state (job_state.JobState)

submit(*, service_account:Optional[str] = None, network:Optional[str] = None, timeout:Optional[int] = None, restart_job_on_worker_restart:bool = False, enable_web_access:bool = False, experiment:Optional[Union[google.cloud.aiplatform.metadata.experiment_resources.Experiment,str]] = None, experiment_run:Optional[Union[google.cloud.aiplatform.metadata.experiment_run_resource.ExperimentRun,str]] = None, tensorboard:Optional[str] = None, create_request_timeout:Optional[float] = None, disable_retries:bool = False, persistent_resource_id:Optional[str] = None, scheduling_strategy:Optional[google.cloud.aiplatform_v1.types.custom_job.Scheduling.Strategy] = None, max_wait_duration:Optional[int] = None, psc_interface_config:Optional[google.cloud.aiplatform_v1.types.service_networking.PscInterfaceConfig] = None)

Submit the configured CustomJob.

• Parameters

  • service_account (str) – Optional. Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.

  • network (str) – Optional. The full name of the Compute Engine network to which the jobshould be peered. For example, projects/12345/global/networks/myVPC.Private services access must already be configured for the network.

  • timeout (int) – The maximum job running time in seconds. The default is 7 days.

  • restart_job_on_worker_restart (bool) – Restarts the entire CustomJob if a workergets restarted. This feature can be used bydistributed training jobs that are not resilientto workers leaving and joining a job.

  • enable_web_access (bool) – Whether you want Vertex AI to enable interactive shell accessto training containers.https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell

  • experiment (Union[aiplatform.Experiment, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The instance or name of an Experiment resource to whichthis CustomJob will upload training parameters and metrics.

    service_account is required with provided experiment.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • experiment_run (Union[aiplatform.ExperimentRun, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The instance or name of an ExperimentRun resource to whichthis CustomJob will upload training parameters and metrics.This arg can only be set when experiment is set. If ‘experiment’is set but ‘experiment_run` is not, an ExperimentRun resourcewill still be auto-generated.

  • tensorboard (str) – Optional. The name of a Vertex AI[Tensorboard][google.cloud.aiplatform.v1beta1.Tensorboard]resource to which this CustomJob will upload Tensorboardlogs. Format:projects/{project}/locations/{location}/tensorboards/{tensorboard}

    The training script should write Tensorboard to following Vertex AI environmentvariable:

    AIP_TENSORBOARD_LOG_DIR

    service_account is required with provided tensorboard.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

  • disable_retries (bool) – Indicates if the job should retry for internal errors after thejob starts running. If True, overridesrestart_job_on_worker_restart to False.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network, CMEK, and node poolconfigs on the job should be consistent with those on thePersistentResource, otherwise, the job will be rejected.

  • scheduling_strategy (gca_custom_job_compat.Scheduling.Strategy) – Optional. Indicates the job scheduling strategy.

  • max_wait_duration (int) – This is the maximum duration that a job will wait for therequested resources to be provisioned in seconds. If set to 0,the job will wait indefinitely. The default is 1 day.

  • psc_interface_config (gca_service_networking.PscInterfaceConfig) – Optional. Configuration for Private Service Connect interfaceused for training.

• Raises

  ValueError – If both experiment and tensorboard are specified or if enable_autolog is True in CustomJob.from_local_script but experiment is not specified or the specified experiment doesn’t have a backing tensorboard.

to_dict()

Returns the resource proto as a dictionary.

property update_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was last updated.

wait()

Helper method that blocks until all futures are complete.

wait_for_completion()

Waits for job to complete.

• Raises

  RuntimeError – If job failed or cancelled.

wait_for_resource_creation()

Waits until resource has been created.

property web_access_uris(: Dict[str, Union[str, Dict[str,str]] )

Fetch the runnable job again and return the latest web access uris.

• Returns

  Web access uris of the runnable job.

• Return type

  (Dict[str, Union[str, Dict[str,str]]])

class google.cloud.aiplatform.CustomPythonPackageTrainingJob(display_name:str, python_package_gcs_uri:Union[str,List[str]], python_module_name:str, container_uri:str, model_serving_container_image_uri:Optional[str] = None, model_serving_container_predict_route:Optional[str] = None, model_serving_container_health_route:Optional[str] = None, model_serving_container_command:Optional[Sequence[str]] = None, model_serving_container_args:Optional[Sequence[str]] = None, model_serving_container_environment_variables:Optional[Dict[str,str]] = None, model_serving_container_ports:Optional[Sequence[int]] = None, model_description:Optional[str] = None, model_instance_schema_uri:Optional[str] = None, model_parameters_schema_uri:Optional[str] = None, model_prediction_schema_uri:Optional[str] = None, explanation_metadata:Optional[google.cloud.aiplatform_v1.types.explanation_metadata.ExplanationMetadata] = None, explanation_parameters:Optional[google.cloud.aiplatform_v1.types.explanation.ExplanationParameters] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None, labels:Optional[Dict[str,str]] = None, training_encryption_spec_key_name:Optional[str] = None, model_encryption_spec_key_name:Optional[str] = None, staging_bucket:Optional[str] = None)

Bases:google.cloud.aiplatform.training_jobs._CustomTrainingJob

Class to launch a Custom Training Job in Vertex AI using a PythonPackage.

Use the CustomPythonPackageTrainingJob class to use a Python package tolaunch a custom training pipeline in Vertex AI. For an example of how to usethe CustomPythonPackageTrainingJob class, see the tutorial in theCustomtraining using Python package, managed text dataset, and TensorFlow servingcontainernotebook.

Constructs a Custom Training Job from a Python Package.

A class to launch a custom training job in Vertex AI using a PythonPackage.

Use the CustomPythonPackageTrainingJob class to use a Python packageto launch a custom training pipeline in Vertex AI. For an example of howto use the CustomPythonPackageTrainingJob class, see the tutorial intheCustom training using Python package, managed text dataset, andTensorFlow servingcontainernotebook.

job = aiplatform.CustomPythonPackageTrainingJob(

display_name=’test-train’,python_package_gcs_uri=’gs://my-bucket/my-python-package.tar.gz’,python_module_name=’my-training-python-package.task’,container_uri=’gcr.io/cloud-aiplatform/training/tf-cpu.2-2:latest’,model_serving_container_image_uri=’gcr.io/my-trainer/serving:1’,model_serving_container_predict_route=’predict’,model_serving_container_health_route=’metadata,labels={‘key’: ‘value’},

)

Usage with Dataset:

ds = aiplatform.TabularDataset(

‘projects/my-project/locations/us-central1/datasets/12345’

)

job.run(

ds,replica_count=1,model_display_name=’my-trained-model’,model_labels={‘key’: ‘value’},

)

Usage without Dataset:

job.run(

replica_count=1,model_display_name=’my-trained-model’,model_labels={‘key’: ‘value’},

)

To ensure your model gets saved in Vertex AI, write your saved model toos.environ[“AIP_MODEL_DIR”] in your provided training script.

• Parameters

  • display_name (str) – Required. The user-defined name of this TrainingPipeline.

  • python_package_gcs_uri (Union[str, **List[str]]) – Required. GCS location of the training python package.Could be a string for single package or a list of string formultiple packages.

  • python_module_name (str) – Required. The module name of the training python package.

  • container_uri (str) – Required. Uri of the training container image in the GCR.

  • model_serving_container_image_uri (str) – Optional. If the training produces a managed Vertex AI Model,the URI of the model serving container suitable for serving themodel produced by the training script.

  • model_serving_container_predict_route (str) – Optional. If the training produces a managed Vertex AI Model,an HTTP path to send prediction requests to the container,and which must be supported by it. If not specified a defaultHTTP path will be used by Vertex AI.

  • model_serving_container_health_route (str) – Optional. If the training produces a managed Vertex AI Model,an HTTP path to send health check requests to the container,and which must be supported by it. If not specified a standardHTTP path will be used by AI Platform.

  • model_serving_container_command (Sequence[str]) – Optional. The command with which the container is run. Not executed within ashell. The Docker image’s ENTRYPOINT is used if this is not provided.Variable references $(VAR_NAME) are expanded using the container’senvironment. If a variable cannot be resolved, the reference in theinput string will be unchanged. The $(VAR_NAME) syntax can be escapedwith a double $$, ie: $$(VAR_NAME). Escaped references will never beexpanded, regardless of whether the variable exists or not.

  • model_serving_container_args (Sequence[str]) – Optional. The arguments to the command. The Docker image’s CMD is used if thisis not provided. Variable references $(VAR_NAME) are expanded using thecontainer’s environment. If a variable cannot be resolved, the referencein the input string will be unchanged. The $(VAR_NAME) syntax can beescaped with a double $$, ie: $$(VAR_NAME). Escaped references willnever be expanded, regardless of whether the variable exists or not.

  • model_serving_container_environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The environment variables that are to be present in the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.

  • model_serving_container_ports (Sequence[int]) – Optional. Declaration of ports that are exposed by the container.This field is primarily informational, it gives Vertex AI informationabout the network connections the container uses. Listing or nota port here has no impact on whether the port is actually exposed,any port listening on the default “0.0.0.0” address inside acontainer will be accessible from the network.

  • model_description (str) – Optional. The description of the Model.

  • model_instance_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the format of a single instance, whichare used inPredictRequest.instances,ExplainRequest.instancesandBatchPredictionJob.input_config.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform. Note: The URI given on output will be immutableand probably different, including the URI scheme, than theone given on input. The output URI will point to a locationwhere the user only has a read access.

  • model_parameters_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the parameters of prediction andexplanation viaPredictRequest.parameters,ExplainRequest.parametersandBatchPredictionJob.model_parameters.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform, if no parameters are supported it is set to anempty string. Note: The URI given on output will beimmutable and probably different, including the URI scheme,than the one given on input. The output URI will point to alocation where the user only has a read access.

  • model_prediction_schema_uri (str) – Optional. Points to a YAML file stored on Google CloudStorage describing the format of a single predictionproduced by this Model, which are returned viaPredictResponse.predictions,ExplainResponse.explanations,andBatchPredictionJob.output_config.The schema is defined as an OpenAPI 3.0.2SchemaObject.AutoML Models always have this field populated by AIPlatform. Note: The URI given on output will be immutableand probably different, including the URI scheme, than theone given on input. The output URI will point to a locationwhere the user only has a read access.

  • explanation_metadata (explain.ExplanationMetadata) – Optional. Metadata describing the Model’s input and output forexplanation. explanation_metadata is optional whileexplanation_parameters must be specified when used.For more details, see Ref docshttp://tinyurl.com/1igh60kt

  • explanation_parameters (explain.ExplanationParameters) – Optional. Parameters to configure explaining for Model’spredictions.For more details, see Ref docshttp://tinyurl.com/1an4zake

  • project (str) – Project to run training in. Overrides project set in aiplatform.init.

  • location (str) – Location to run training in. Overrides location set in aiplatform.init.

  • credentials (auth_credentials.Credentials) – Custom credentials to use to run call training service. Overridescredentials set in aiplatform.init.

  • labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize TrainingPipelines.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • training_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the training pipeline. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, this TrainingPipeline will be secured by this key.

    Note: Model trained by this TrainingPipeline is also securedby this key ifmodel_to_upload is not set separately.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • model_encryption_spec_key_name (Optional[str]) – Optional. The Cloud KMS resource identifier of the customermanaged encryption key used to protect the model. Has theform:projects/my-project/locations/my-region/keyRings/my-kr/cryptoKeys/my-key.The key needs to be in the same region as where the computeresource is created.

    If set, the trained Model will be secured by this key.

    Overrides encryption_spec_key_name set in aiplatform.init.

  • staging_bucket (str) – Optional. Bucket used to stage source and training artifacts. Overridesstaging_bucket set in aiplatform.init.

cancel()

Asynchronously attempts to cancel a training job.

The server makes a best effort to cancel the job, but the training jobcan’t always be cancelled. If the training job is canceled, its statetransitions to CANCELLED and it’s not deleted.

• Raises

  RuntimeError – If this training job isn’t running, then a runtime error is raised.

client_class()

alias ofgoogle.cloud.aiplatform.utils.PipelineClientWithOverride

property create_time(: [datetime.datetime](https://docs.python.org/3/library/datetime.html#datetime.datetime )

Time this resource was created.

delete(sync:bool = True)

Deletes this Vertex AI resource. WARNING: This deletion ispermanent.

• Parameters

  sync (bool) – Whether to execute this deletion synchronously. If False, this method will be executed in concurrent Future and any downstream object will be immediately returned and synced when the Future has completed.

property display_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Display name of this resource.

done()

Method indicating whether a job has completed.

• Returns

  True if the job has completed.

property encryption_spec(: Optional[[google.cloud.aiplatform_v1.types.encryption_spec.EncryptionSpec](../aiplatform_v1/types.md#google.cloud.aiplatform_v1.types.EncryptionSpec)_ )

Customer-managed encryption key options for this Vertex AI resource.

If this is set, then all resources created by this Vertex AI resource willbe encrypted with the provided encryption key.

property end_time(: Optional[datetime.datetime )

Optional. The time when the training job entered thePIPELINE_STATE_SUCCEEDED, PIPELINE_STATE_FAILED, orPIPELINE_STATE_CANCELLED state.

property error(: Optional[google.rpc.status_pb2.Status )

Optional. Detailed error information for this training job resource.Error information is created only when the state of the training job isPIPELINE_STATE_FAILED or PIPELINE_STATE_CANCELLED.

property gca_resource(: [proto.message.Message](https://proto-plus-python.readthedocs.io/en/latest/reference/message.html#proto.message.Message )

The underlying resource proto representation.

classmethod get(resource_name:str, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Gets a training job using the resource_name that’s passed in.

• Parameters

  • resource_name (str) – Required. A fully-qualified resource name or ID.

  • project (str) – Optional. The name of the Google Cloud project to retrieve thetraining job from. This overrides the project that was set byaiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training job isretrieved. This region overrides the region that was set byaiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to upload this model.These credentials override the credentials set byaiplatform.init.

• Raises

  ValueError – A ValueError is raised if the task definition of the retrieved training job doesn’t match the custom training task definition.

• Returns

  A Vertex AI training job.

get_model(sync=True)

Returns the Vertex AI model produced by this training job.

• Parameters

  sync (bool) – If set to true, this method runs synchronously. If false, this method runs asynchronously.

• Returns

  The Vertex AI model that was produced by this training job.

• Raises

  RuntimeError – A runtime error is raised if the training job failed or if a model wasn’t produced by the training job.

property has_failed(: [bool](https://docs.python.org/3/library/functions.html#bool )

Returns true if the training job failed, otherwise false.

property labels(: Dict[str,str )

User-defined labels containing metadata about this resource.

Read more about labels athttps://goo.gl/xmQnxf

classmethod list(filter:Optional[str] = None, order_by:Optional[str] = None, project:Optional[str] = None, location:Optional[str] = None, credentials:Optional[google.auth.credentials.Credentials] = None)

Lists all instances of this training job resource.

The following shows an example of how to call CustomTrainingJob.list:

``

`

pyaiplatform.CustomTrainingJob.list(

filter=’display_name=”experiment_a27”’,order_by=’create_time desc’

• Parameters

  • filter (str) – Optional. An expression for filtering the results of the request.For field names, snake_case and camelCase are supported.

  • order_by (str) – Optional. A comma-separated list of fields used to sort thereturned traing job resources. The defauilt sorting order isascending. To sort by a field name in descending order, usedesc after the field name. The following fields are supported:display_name, create_time, update_time.

  • project (str) – Optional. The name of the Google Cloud project to which toretrieve the list of training job resources. This overrides theproject that was set by aiplatform.init.

  • location (str) – Optional. The Google Cloud region from where the training jobresources are retrieved. This region overrides the region thatwas set by aiplatform.init.

  • credentials (auth_credentials.Credentials) – Optional. The credentials that are used to retrieve list. Thesecredentials override the credentials set by aiplatform.init.

• Returns

  A list of training job resources.

• Return type

  List[VertexAiResourceNoun]

property name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Name of this resource.

property network(: Optional[str )

The full name of the Google Compute Enginenetwork to which thisCustomTrainingJob should be peered.

Specify the name of the network using the formatprojects/{project}/global/networks/{network}. Replace {project} withthe project number, such as 12345, and {network} with a network name.

Before specifying a network, private services access must be configuredfor the network. If private services access isn’t configured, thenthe custom training job can’t be peered with a network.

property resource_name(: [str](https://docs.python.org/3/library/stdtypes.html#str )

Full qualified resource name.

run(dataset:Optional[Union[google.cloud.aiplatform.datasets.image_dataset.ImageDataset, google.cloud.aiplatform.datasets.tabular_dataset.TabularDataset, google.cloud.aiplatform.datasets.text_dataset.TextDataset, google.cloud.aiplatform.datasets.video_dataset.VideoDataset]] = None, annotation_schema_uri:Optional[str] = None, model_display_name:Optional[str] = None, model_labels:Optional[Dict[str,str]] = None, model_id:Optional[str] = None, parent_model:Optional[str] = None, is_default_version:Optional[bool] = True, model_version_aliases:Optional[Sequence[str]] = None, model_version_description:Optional[str] = None, base_output_dir:Optional[str] = None, service_account:Optional[str] = None, network:Optional[str] = None, bigquery_destination:Optional[str] = None, args:Optional[List[Union[str,float,int]]] = None, environment_variables:Optional[Dict[str,str]] = None, replica_count:int = 1, machine_type:str = 'n1-standard-4', accelerator_type:str = 'ACCELERATOR_TYPE_UNSPECIFIED', accelerator_count:int = 0, boot_disk_type:str = 'pd-ssd', boot_disk_size_gb:int = 100, reduction_server_replica_count:int = 0, reduction_server_machine_type:Optional[str] = None, reduction_server_container_uri:Optional[str] = None, training_fraction_split:Optional[float] = None, validation_fraction_split:Optional[float] = None, test_fraction_split:Optional[float] = None, training_filter_split:Optional[str] = None, validation_filter_split:Optional[str] = None, test_filter_split:Optional[str] = None, predefined_split_column_name:Optional[str] = None, timestamp_split_column_name:Optional[str] = None, timeout:Optional[int] = None, restart_job_on_worker_restart:bool = False, enable_web_access:bool = False, enable_dashboard_access:bool = False, tensorboard:Optional[str] = None, sync=True, create_request_timeout:Optional[float] = None, disable_retries:bool = False, persistent_resource_id:Optional[str] = None, tpu_topology:Optional[str] = None, scheduling_strategy:Optional[google.cloud.aiplatform_v1.types.custom_job.Scheduling.Strategy] = None, reservation_affinity_type:Optional[Literal['NO_RESERVATION', 'ANY_RESERVATION', 'SPECIFIC_RESERVATION']] = None, reservation_affinity_key:Optional[str] = None, reservation_affinity_values:Optional[List[str]] = None, max_wait_duration:Optional[int] = None, psc_interface_config:Optional[google.cloud.aiplatform_v1.types.service_networking.PscInterfaceConfig] = None)

Runs the custom training job.

Distributed Training Support:If replica count = 1 then one chief replica will be provisioned. Ifreplica_count > 1 the remainder will be provisioned as a worker replica pool.ie: replica_count = 10 will result in 1 chief and 9 workersAll replicas have same machine_type, accelerator_type, and accelerator_count

If training on a Vertex AI dataset, you can use one of the following split configurations:

Data fraction splits:Any of `training_fraction_split`, `validation_fraction_split` and`test_fraction_split` may optionally be provided, they must sum to up to 1. Ifthe provided ones sum to less than 1, the remainder is assigned to sets asdecided by Vertex AI. If none of the fractions are set, by default roughly 80%of data will be used for training, 10% for validation, and 10% for test.Data filter splits:Assigns input data to training, validation, and test setsbased on the given filters, data pieces not matched by anyfilter are ignored. Currently only supported for Datasetscontaining DataItems.If any of the filters in this message are to match nothing, thenthey can be set as ‘-’ (the minus sign).If using filter splits, all of `training_filter_split`, `validation_filter_split` and`test_filter_split` must be provided.Supported only for unstructured Datasets.Predefined splits:Assigns input data to training, validation, and test sets based on the value of a provided key.If using predefined splits, `predefined_split_column_name` must be provided.Supported only for tabular Datasets.Timestamp splits:Assigns input data to training, validation, and test setsbased on a provided timestamps. The youngest data pieces areassigned to training set, next to validation set, and the oldestto the test set.Supported only for tabular Datasets.

• Parameters

  • dataset (Union[datasets.ImageDataset,datasets.TabularDataset,datasets.TextDataset,datasets.VideoDataset,]) – Vertex AI to fit this training against. Custom training script shouldretrieve datasets through passed in environment variables uris:

    os.environ[“AIP_TRAINING_DATA_URI”]os.environ[“AIP_VALIDATION_DATA_URI”]os.environ[“AIP_TEST_DATA_URI”]

    Additionally the dataset format is passed in as:

    os.environ[“AIP_DATA_FORMAT”]

  • annotation_schema_uri (str) – Google Cloud Storage URI points to a YAML file describingannotation schema. The schema is defined as an OpenAPI 3.0.2Schema Object The schema filesthat can be used here are found ings://google-cloud-aiplatform/schema/dataset/annotation/,note that the chosen schema must be consistent withmetadataof the Dataset specified bydataset_id.

    Only Annotations that both match this schema and belong toDataItems not ignored by the split method are used inrespectively training, validation or test role, depending onthe role of the DataItem they are on.

    When used in conjunction withannotations_filter,the Annotations used for training are filtered by bothannotations_filterandannotation_schema_uri.

  • model_display_name (str) – If the script produces a managed Vertex AI Model. The display name ofthe Model. The name can be up to 128 characters long and can be consistof any UTF-8 characters.

    If not provided upon creation, the job’s display_name is used.

  • model_labels (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Optional. The labels with user-defined metadata toorganize your Models.Label keys and values can be no longer than 64characters (Unicode codepoints), can onlycontain lowercase letters, numeric characters,underscores and dashes. International charactersare allowed.Seehttps://goo.gl/xmQnxf for more informationand examples of labels.

  • model_id (str) – Optional. The ID to use for the Model produced by this job,which will become the final component of the model resource name.This value may be up to 63 characters, and valid charactersare [a-z0-9_-]. The first character cannot be a number or hyphen.

  • parent_model (str) – Optional. The resource name or model ID of an existing model.The new model uploaded by this job will be a version of parent_model.

    Only set this field when training a new version of an existing model.

  • is_default_version (bool) – Optional. When set to True, the newly uploaded model version willautomatically have alias “default” included. Subsequent uses ofthe model produced by this job without a version specified willuse this “default” version.

    When set to False, the “default” alias will not be moved.Actions targeting the model version produced by this job will needto specifically reference this version by ID or alias.

    New model uploads, i.e. version 1, will always be “default” aliased.

  • model_version_aliases (Sequence[str]) – Optional. User provided version aliases so that the model versionuploaded by this job can be referenced via alias instead ofauto-generated version ID. A default version alias will be createdfor the first version of the model.

    The format is [a-z][a-zA-Z0-9-]{0,126}[a-z0-9]

  • model_version_description (str) – Optional. The description of the model version being uploaded by this job.

  • base_output_dir (str) – GCS output directory of job. If not provided atimestamped directory in the staging directory will be used.

    Vertex AI sets the following environment variables when it runs your training code:

    • AIP_MODEL_DIR: a Cloud Storage URI of a directory intended for saving model artifacts, i.e. <base_output_dir>/model/

    • AIP_CHECKPOINT_DIR: a Cloud Storage URI of a directory intended for saving checkpoints, i.e. <base_output_dir>/checkpoints/

    • AIP_TENSORBOARD_LOG_DIR: a Cloud Storage URI of a directory intended for saving TensorBoard logs, i.e. <base_output_dir>/logs/

  • service_account (str) – Specifies the service account for workload run-as account.Users submitting jobs must have act-as permission on this run-as account.If not specified, uses the service account set in aiplatform.init.

  • network (str) – The full name of the Compute Engine network to which the jobshould be peered. For example, projects/12345/global/networks/myVPC.Private services access must already be configured for the network.If left unspecified, the network set in aiplatform.init will be used.Otherwise, the job is not peered with any network.

  • bigquery_destination (str) – Provide this field if dataset is a BigQuery dataset.The BigQuery project location where the training data is tobe written to. In the given project a new dataset is createdwith namedataset_<dataset-id>_<annotation-type>_<timestamp-of-training-call>where timestamp is in YYYY_MM_DDThh_mm_ss_sssZ format. Alltraining input data will be written into that dataset. Inthe dataset three tables will be created,training,validation andtest.

    • AIP_DATA_FORMAT = “bigquery”.

    • AIP_TRAINING_DATA_URI =”bigquery_destination.dataset_*.training”

    • AIP_VALIDATION_DATA_URI = “bigquery_destination.dataset_*.validation”

    • AIP_TEST_DATA_URI = “bigquery_destination.dataset_*.test”

  • args (List[Unions[str, *[int](https://docs.python.org/3/library/functions.html#int),[float](https://docs.python.org/3/library/functions.html#float)]*]) – Command line arguments to be passed to the Python script.

  • environment_variables (Dict[str, *[str](https://docs.python.org/3/library/stdtypes.html#str)]*) – Environment variables to be passed to the container.Should be a dictionary where keys are environment variable namesand values are environment variable values for those names.At most 10 environment variables can be specified.The Name of the environment variable must be unique.

    environment_variables = {

    ‘MY_KEY’: ‘MY_VALUE’

    }

  • replica_count (int) – The number of worker replicas. If replica count = 1 then one chiefreplica will be provisioned. If replica_count > 1 the remainder will beprovisioned as a worker replica pool.

  • machine_type (str) – The type of machine to use for training.

  • accelerator_type (str) – Hardware accelerator type. One of ACCELERATOR_TYPE_UNSPECIFIED,NVIDIA_TESLA_K80, NVIDIA_TESLA_P100, NVIDIA_TESLA_V100, NVIDIA_TESLA_P4,NVIDIA_TESLA_T4

  • accelerator_count (int) – The number of accelerators to attach to a worker replica.

  • boot_disk_type (str) – Type of the boot disk, default is pd-ssd.Valid values: pd-ssd (Persistent Disk Solid State Drive) orpd-standard (Persistent Disk Hard Disk Drive).

  • boot_disk_size_gb (int) – Size in GB of the boot disk, default is 100GB.boot disk size must be within the range of [100, 64000].

  • reduction_server_replica_count (int) – The number of reduction server replicas, default is 0.

  • reduction_server_machine_type (str) – Optional. The type of machine to use for reduction server.

  • reduction_server_container_uri (str) – Optional. The Uri of the reduction server container image.See details:https://cloud.google.com/vertex-ai/docs/training/distributed-training#reduce_training_time_with_reduction_server

  • training_fraction_split (float) – Optional. The fraction of the input data that is to be used to trainthe Model. This is ignored if Dataset is not provided.

  • validation_fraction_split (float) – Optional. The fraction of the input data that is to be used to validatethe Model. This is ignored if Dataset is not provided.

  • test_fraction_split (float) – Optional. The fraction of the input data that is to be used to evaluatethe Model. This is ignored if Dataset is not provided.

  • training_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to train the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • validation_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to validate the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • test_filter_split (str) – Optional. A filter on DataItems of the Dataset. DataItems that matchthis filter are used to test the Model. A filter with same syntaxas the one used in DatasetService.ListDataItems may be used. If asingle DataItem is matched by more than one of the FilterSplit filters,then it is assigned to the first set that applies to it in the training,validation, test order. This is ignored if Dataset is not provided.

  • predefined_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key (either the label’s value orvalue in the column) must be one of {training,validation,test}, and it defines to which set thegiven piece of data is assigned. If for a piece of data thekey is not present or has an invalid value, that piece isignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timestamp_split_column_name (str) – Optional. The key is a name of one of the Dataset’s datacolumns. The value of the key values of the key (the values inthe column) must be in RFC 3339 date-time format, wheretime-offset = “Z” (e.g. 1985-04-12T23:20:50.52Z). If for apiece of data the key is not present or has an invalid value,that piece is ignored by the pipeline.

    Supported only for tabular and time series Datasets.

  • timeout (int) – The maximum job running time in seconds. The default is 7 days.

  • restart_job_on_worker_restart (bool) – Restarts the entire CustomJob if a workergets restarted. This feature can be used bydistributed training jobs that are not resilientto workers leaving and joining a job.

  • enable_web_access (bool) – Whether you want Vertex AI to enable interactive shell accessto training containers.https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell

  • enable_dashboard_access (bool) – Whether you want Vertex AI to enable access to the customized dashboardto training containers.

  • tensorboard (str) – Optional. The name of a Vertex AI[Tensorboard][google.cloud.aiplatform.v1beta1.Tensorboard]resource to which this CustomJob will upload Tensorboardlogs. Format:projects/{project}/locations/{location}/tensorboards/{tensorboard}

    The training script should write Tensorboard to following Vertex AI environmentvariable:

    AIP_TENSORBOARD_LOG_DIR

    service_account is required with provided tensorboard.For more information on configuring your service account please visit:https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-training

  • sync (bool) – Whether to execute this method synchronously. If False, this methodwill be executed in concurrent Future and any downstream object willbe immediately returned and synced when the Future has completed.

  • create_request_timeout (float) – Optional. The timeout for the create request in seconds.

  • disable_retries (bool) – Indicates if the job should retry for internal errors after thejob starts running. If True, overridesrestart_job_on_worker_restart to False.

  • persistent_resource_id (str) – Optional. The ID of the PersistentResource in the same Projectand Location. If this is specified, the job will be run onexisting machines held by the PersistentResource instead ofon-demand short-live machines. The network, CMEK, and node poolconfigs on the job should be consistent with those on thePersistentResource, otherwise, the job will be rejected.

  • tpu_topology (str) – Optional. Specifies the tpu topology to be used forTPU training job. This field is required for TPU v5 versions. Fordetails on the TPU topology, refer tohttps://cloud.google.com/tpu/docs/v5e#tpu-v5e-config. The topologymust be a supported value for the TPU machine type.

  • scheduling_strategy (gca_custom_job_compat.Scheduling.Strategy) – Optional. Indicates the job scheduling strategy.

  • reservation_affinity_type (str) – Optional. The type of reservation affinity. One of:* “NO_RESERVATION” : No reservation is used.* “ANY_RESERVATION” : Any reservation that matches machine speccan be used.* “SPECIFIC_RESERVATION” : A specific reservation must be useused. See reservation_affinity_key andreservation_affinity_values for how to specify the reservation.

  • reservation_affinity_key (str) – Optional. Corresponds to the label key of a reservation resource.To target a SPECIFIC_RESERVATION by name, usecompute.googleapis.com/reservation-name as the keyand specify the name of your reservation as its value.

  • reservation_affinity_values (List[str]) – Optional. Corresponds to the label values of a reservation resource.This must be the full resource name of the reservation.Format: ‘projects/{project_id_or_number}/zones/{zone}/reservations/{reservation_name}’

  • max_wait_duration (int) – This is the maximum duration that a job will wait for therequested resources to be provisioned in seconds. If set to 0,the job will wait indefinitely. The default is 30 minutes.

  • psc_interface_config (gca_service_networking.PscInterfaceConfig) – Optional. Configuration for Private Service Connect interfaceused for training.

• Returns

  The trained Vertex AI Model resource or None if training did not

    produce a Vertex AI Model.

• Return type

  model