class SageMakerWorkteamSpec(SageMakerComponentSpec[SageMakerWorkteamInputs,
SageMakerWorkteamOutputs]):
INPUTS: SageMakerWorkteamInputs = SageMakerWorkteamInputs(
team_name=InputValidator(input_type=str,
required=True,
description="The name of your work team."),
description=InputValidator(
input_type=str,
required=True,
description="A description of the work team."),
user_pool=InputValidator(
input_type=str,
required=False,
description=
"An identifier for a user pool. The user pool must be in the same region as the service that you are calling.",
),
user_groups=InputValidator(
input_type=str,
required=False,
description=
"A list of identifiers for user groups separated by commas.",
default="",
),
client_id=InputValidator(
input_type=str,
required=False,
description=
"An identifier for an application client. You must create the app client ID using Amazon Cognito.",
),
sns_topic=InputValidator(
input_type=str,
required=False,
description=
"The ARN for the SNS topic to which notifications should be published.",
default="",
),
**vars(COMMON_INPUTS),
)
OUTPUTS = SageMakerWorkteamOutputs(
workteam_arn=OutputValidator(description="The ARN of the workteam."), )
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerWorkteamInputs,
SageMakerWorkteamOutputs)
@property
def inputs(self) -> SageMakerWorkteamInputs:
return self._inputs
@property
def outputs(self) -> SageMakerWorkteamOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerWorkteamOutputs:
return self._output_paths
class DummySpec(SageMakerComponentSpec[DummyInputs, DummyOutputs]):
INPUTS: DummyInputs = DummyInputs(
input1=InputValidator(
input_type=str,
description="The first input.",
default="input1-default",
),
input2=InputValidator(
input_type=int,
required=True,
description="The second input.",
),
)
OUTPUTS = DummyOutputs(
output1=OutputValidator(description="The first output."),
output2=OutputValidator(description="The second output."),
)
def __init__(self, arguments: List[str]):
super().__init__(arguments, DummyInputs, DummyOutputs)
class RoboMakerDeleteSimulationAppSpec(
SageMakerComponentSpec[RoboMakerDeleteSimulationAppInputs,
RoboMakerDeleteSimulationAppOutputs]):
INPUTS: RoboMakerDeleteSimulationAppInputs = RoboMakerDeleteSimulationAppInputs(
arn=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) of the simulation application.",
default="",
),
version=InputValidator(
input_type=str,
required=False,
description="The version of the simulation application.",
default=None,
),
**vars(COMMON_INPUTS),
)
OUTPUTS = RoboMakerDeleteSimulationAppOutputs(arn=OutputValidator(
description=
"The Amazon Resource Name (ARN) of the simulation application."), )
def __init__(self, arguments: List[str]):
super().__init__(
arguments,
RoboMakerDeleteSimulationAppInputs,
RoboMakerDeleteSimulationAppOutputs,
)
@property
def inputs(self) -> RoboMakerDeleteSimulationAppInputs:
return self._inputs
@property
def outputs(self) -> RoboMakerDeleteSimulationAppOutputs:
return self._outputs
@property
def output_paths(self) -> RoboMakerDeleteSimulationAppOutputs:
return self._output_paths
class RoboMakerSimulationJobBatchSpec(
SageMakerComponentSpec[RoboMakerSimulationJobBatchInputs,
RoboMakerSimulationJobBatchOutputs]):
INPUTS: RoboMakerSimulationJobBatchInputs = RoboMakerSimulationJobBatchInputs(
role=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) that Amazon RoboMaker assumes to perform tasks on your behalf.",
),
timeout_in_secs=InputValidator(
input_type=int,
required=False,
description=
"The amount of time, in seconds, to wait for the batch to complete.",
default=0,
),
max_concurrency=InputValidator(
input_type=int,
required=False,
description=
"The number of active simulation jobs create as part of the batch that can be in an active state at the same time.",
default=0,
),
simulation_job_requests=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=True,
description=
"A list of simulation job requests to create in the batch.",
default=[],
),
sim_app_arn=InputValidator(
input_type=str,
required=False,
description="The application ARN for the simulation application.",
default="",
),
**vars(COMMON_INPUTS),
)
OUTPUTS = RoboMakerSimulationJobBatchOutputs(
arn=OutputValidator(
description="The Amazon Resource Name (ARN) of the simulation job."
),
batch_job_id=OutputValidator(
description="The simulation job batch id."),
)
def __init__(self, arguments: List[str]):
super().__init__(
arguments,
RoboMakerSimulationJobBatchInputs,
RoboMakerSimulationJobBatchOutputs,
)
@property
def inputs(self) -> RoboMakerSimulationJobBatchInputs:
return self._inputs
@property
def outputs(self) -> RoboMakerSimulationJobBatchOutputs:
return self._outputs
@property
def output_paths(self) -> RoboMakerSimulationJobBatchOutputs:
return self._output_paths
class SageMakerTuningSpec(SageMakerComponentSpec[SageMakerTuningInputs,
SageMakerTuningOutputs]):
INPUTS: SageMakerTuningInputs = SageMakerTuningInputs(
job_name=InputValidator(
input_type=str,
required=False,
description=
"The name of the tuning job. Must be unique within the same AWS account and AWS region.",
),
role=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) that Amazon SageMaker assumes to perform tasks on your behalf.",
),
image=InputValidator(
input_type=str,
required=False,
description=
"The registry path of the Docker image that contains the training algorithm.",
default="",
),
algorithm_name=InputValidator(
input_type=str,
required=False,
description=
"The name of the resource algorithm to use for the hyperparameter tuning job.",
default="",
),
training_input_mode=InputValidator(
choices=["File", "Pipe"],
input_type=str,
required=False,
description=
"The input mode that the algorithm supports. File or Pipe.",
default="File",
),
metric_definitions=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"The dictionary of name-regex pairs specify the metrics that the algorithm emits.",
default={},
),
strategy=InputValidator(
choices=["Bayesian", "Random"],
input_type=str,
required=False,
description=
"How hyperparameter tuning chooses the combinations of hyperparameter values to use for the training job it launches.",
default="Bayesian",
),
metric_name=InputValidator(
input_type=str,
required=True,
description=
"The name of the metric to use for the objective metric.",
),
metric_type=InputValidator(
choices=["Maximize", "Minimize"],
input_type=str,
required=True,
description="Whether to minimize or maximize the objective metric.",
),
early_stopping_type=InputValidator(
choices=["Off", "Auto"],
input_type=str,
required=False,
description="Whether to minimize or maximize the objective metric.",
default="Off",
),
static_parameters=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"The values of hyperparameters that do not change for the tuning job.",
default={},
),
integer_parameters=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The array of IntegerParameterRange objects that specify ranges of integer hyperparameters that you want to search.",
default=[],
),
continuous_parameters=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The array of ContinuousParameterRange objects that specify ranges of continuous hyperparameters that you want to search.",
default=[],
),
categorical_parameters=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The array of CategoricalParameterRange objects that specify ranges of categorical hyperparameters that you want to search.",
default=[],
),
channels=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=True,
description=
"A list of dicts specifying the input channels. Must have at least one.",
),
output_location=InputValidator(
input_type=str,
required=True,
description=
"The Amazon S3 path where you want Amazon SageMaker to store the results of the transform job.",
),
output_encryption_key=InputValidator(
input_type=str,
required=False,
description=
"The AWS KMS key that Amazon SageMaker uses to encrypt the model artifacts.",
default="",
),
instance_type=InputValidator(
input_type=str,
required=False,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
instance_count=InputValidator(
input_type=int,
required=False,
description=
"The number of ML compute instances to use in each training job.",
default=1,
),
volume_size=InputValidator(
input_type=int,
required=False,
description=
"The size of the ML storage volume that you want to provision.",
default=30,
),
max_num_jobs=InputValidator(
input_type=int,
required=True,
description=
"The maximum number of training jobs that a hyperparameter tuning job can launch.",
),
max_parallel_jobs=InputValidator(
input_type=int,
required=True,
description=
"The maximum number of concurrent training jobs that a hyperparameter tuning job can launch.",
),
resource_encryption_key=InputValidator(
input_type=str,
required=False,
description=
"The AWS KMS key that Amazon SageMaker uses to encrypt data on the storage volume attached to the ML compute instance(s).",
default="",
),
vpc_security_group_ids=InputValidator(
input_type=str,
required=False,
description="The VPC security group IDs, in the form sg-xxxxxxxx.",
),
vpc_subnets=InputValidator(
input_type=str,
required=False,
description=
"The ID of the subnets in the VPC to which you want to connect your hpo job.",
),
network_isolation=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="Isolates the training container.",
default=True,
),
traffic_encryption=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description=
"Encrypts all communications between ML compute instances in distributed training.",
default=False,
),
warm_start_type=InputValidator(
choices=["IdenticalDataAndAlgorithm", "TransferLearning", ""],
input_type=str,
required=False,
description=
'Specifies either "IdenticalDataAndAlgorithm" or "TransferLearning"',
),
parent_hpo_jobs=InputValidator(
input_type=str,
required=False,
description=
"List of previously completed or stopped hyperparameter tuning jobs to be used as a starting point.",
default="",
),
**vars(COMMON_INPUTS),
**vars(SPOT_INSTANCE_INPUTS))
OUTPUTS = SageMakerTuningOutputs(
hpo_job_name=OutputValidator(
description="The name of the hyper parameter tuning job."),
model_artifact_url=OutputValidator(
description="The output model artifacts S3 url."),
best_job_name=OutputValidator(
description="Best training job in the hyper parameter tuning job."
),
best_hyperparameters=OutputValidator(
description="The resulting tuned hyperparameters."),
training_image=OutputValidator(
description=
"The registry path of the Docker image that contains the training algorithm."
),
)
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerTuningInputs,
SageMakerTuningOutputs)
@property
def inputs(self) -> SageMakerTuningInputs:
return self._inputs
@property
def outputs(self) -> SageMakerTuningOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerTuningOutputs:
return self._output_paths
class SageMakerGroundTruthSpec(
SageMakerComponentSpec[SageMakerGroundTruthInputs, SageMakerGroundTruthOutputs]
):
INPUTS: SageMakerGroundTruthInputs = SageMakerGroundTruthInputs(
role=InputValidator(
input_type=str,
required=True,
description="The Amazon Resource Name (ARN) that Amazon SageMaker assumes to perform tasks on your behalf.",
),
job_name=InputValidator(
input_type=str, description="The name of the labeling job."
),
label_attribute_name=InputValidator(
input_type=str,
required=False,
description="The attribute name to use for the label in the output manifest file. Default is the job name.",
default="",
),
manifest_location=InputValidator(
input_type=str,
required=True,
description="The Amazon S3 location of the manifest file that describes the input data objects.",
),
output_location=InputValidator(
input_type=str,
required=True,
description="The Amazon S3 location to write output data.",
),
output_encryption_key=InputValidator(
input_type=str,
required=False,
description="The AWS KMS key that Amazon SageMaker uses to encrypt the model artifacts.",
default="",
),
task_type=InputValidator(
input_type=str,
required=True,
description="Built in image classification, bounding box, text classification, or semantic segmentation, or custom. If custom, please provide pre- and post-labeling task lambda functions.",
),
worker_type=InputValidator(
input_type=str,
required=True,
description="The workteam for data labeling, either public, private, or vendor.",
),
workteam_arn=InputValidator(
input_type=str,
required=False,
description="The ARN of the work team assigned to complete the tasks.",
),
no_adult_content=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="If true, your data is free of adult content.",
default="False",
),
no_ppi=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="If true, your data is free of personally identifiable information.",
default="False",
),
label_category_config=InputValidator(
input_type=str,
required=False,
description="The S3 URL of the JSON structured file that defines the categories used to label the data objects.",
default="",
),
max_human_labeled_objects=InputValidator(
input_type=int,
required=False,
description="The maximum number of objects that can be labeled by human workers.",
default=0,
),
max_percent_objects=InputValidator(
input_type=int,
required=False,
description="The maximum percentatge of input data objects that should be labeled.",
default=0,
),
enable_auto_labeling=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="Enables auto-labeling, only for bounding box, text classification, and image classification.",
default=False,
),
initial_model_arn=InputValidator(
input_type=str,
required=False,
description="The ARN of the final model used for a previous auto-labeling job.",
default="",
),
resource_encryption_key=InputValidator(
input_type=str,
required=False,
description="The AWS KMS key that Amazon SageMaker uses to encrypt data on the storage volume attached to the ML compute instance(s).",
default="",
),
ui_template=InputValidator(
input_type=str,
required=True,
description="The Amazon S3 bucket location of the UI template.",
),
pre_human_task_function=InputValidator(
input_type=str,
required=False,
description="The ARN of a Lambda function that is run before a data object is sent to a human worker.",
default="",
),
post_human_task_function=InputValidator(
input_type=str,
required=False,
description="The ARN of a Lambda function implements the logic for annotation consolidation.",
default="",
),
task_keywords=InputValidator(
input_type=str,
required=False,
description="Keywords used to describe the task so that workers on Amazon Mechanical Turk can discover the task.",
default="",
),
title=InputValidator(
input_type=str,
required=True,
description="A title for the task for your human workers.",
),
description=InputValidator(
input_type=str,
required=True,
description="A description of the task for your human workers.",
),
num_workers_per_object=InputValidator(
input_type=int,
required=True,
description="The number of human workers that will label an object.",
),
time_limit=InputValidator(
input_type=int,
required=True,
description="The amount of time that a worker has to complete a task in seconds",
),
task_availibility=InputValidator(
input_type=int,
required=False,
description="The length of time that a task remains available for labelling by human workers.",
default="0",
),
max_concurrent_tasks=InputValidator(
input_type=int,
required=False,
description="The maximum number of data objects that can be labeled by human workers at the same time.",
default="0",
),
workforce_task_price=InputValidator(
input_type=float,
required=False,
description="The price that you pay for each task performed by a public worker in USD. Specify to the tenth fractions of a cent. Format as '0.000'.",
default="0.000",
),
**vars(COMMON_INPUTS),
)
OUTPUTS = SageMakerGroundTruthOutputs(
output_manifest_location=OutputValidator(
description="The Amazon S3 bucket location of the manifest file for labeled data."
),
active_learning_model_arn=OutputValidator(
description="The ARN for the most recent Amazon SageMaker model trained as part of automated data labeling."
),
)
def __init__(self, arguments: List[str]):
super().__init__(
arguments, SageMakerGroundTruthInputs, SageMakerGroundTruthOutputs
)
@property
def inputs(self) -> SageMakerGroundTruthInputs:
return self._inputs
@property
def outputs(self) -> SageMakerGroundTruthOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerGroundTruthOutputs:
return self._output_paths
class SageMakerDeploySpec(SageMakerComponentSpec[SageMakerDeployInputs,
SageMakerDeployOutputs]):
INPUTS: SageMakerDeployInputs = SageMakerDeployInputs(
endpoint_config_name=InputValidator(
input_type=str,
required=False,
description=
"The name of the endpoint configuration. If an existing endpoint is being updated, a suffix is automatically added if this config name exists.",
default="",
),
update_endpoint=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="Update endpoint if it exists.",
default=False,
),
variant_name_1=InputValidator(
input_type=str,
required=False,
description="The name of the production variant.",
default="variant-name-1",
),
model_name_1=InputValidator(
input_type=str,
required=True,
description="The model name used for endpoint deployment.",
),
initial_instance_count_1=InputValidator(
input_type=int,
required=False,
description="Number of instances to launch initially.",
default=1,
),
instance_type_1=InputValidator(
input_type=str,
required=False,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
initial_variant_weight_1=InputValidator(
input_type=float,
required=False,
description=
"Determines initial traffic distribution among all of the models that you specify in the endpoint configuration.",
default=1.0,
),
accelerator_type_1=InputValidator(
choices=["ml.eia1.medium", "ml.eia1.large", "ml.eia1.xlarge", ""],
input_type=str,
required=False,
description=
"The size of the Elastic Inference (EI) instance to use for the production variant.",
default="",
),
variant_name_2=InputValidator(
input_type=str,
required=False,
description="The name of the production variant.",
default="variant-name-2",
),
model_name_2=InputValidator(
input_type=str,
required=False,
description="The model name used for endpoint deployment.",
default="",
),
initial_instance_count_2=InputValidator(
input_type=int,
required=False,
description="Number of instances to launch initially.",
default=1,
),
instance_type_2=InputValidator(
input_type=str,
required=False,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
initial_variant_weight_2=InputValidator(
input_type=float,
required=False,
description=
"Determines initial traffic distribution among all of the models that you specify in the endpoint configuration.",
default=1.0,
),
accelerator_type_2=InputValidator(
choices=["ml.eia1.medium", "ml.eia1.large", "ml.eia1.xlarge", ""],
input_type=str,
required=False,
description=
"The size of the Elastic Inference (EI) instance to use for the production variant.",
default="",
),
variant_name_3=InputValidator(
input_type=str,
required=False,
description="The name of the production variant.",
default="variant-name-3",
),
model_name_3=InputValidator(
input_type=str,
required=False,
description="The model name used for endpoint deployment.",
default="",
),
initial_instance_count_3=InputValidator(
input_type=int,
required=False,
description="Number of instances to launch initially.",
default=1,
),
instance_type_3=InputValidator(
input_type=str,
required=False,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
initial_variant_weight_3=InputValidator(
input_type=float,
required=False,
description=
"Determines initial traffic distribution among all of the models that you specify in the endpoint configuration.",
default=1.0,
),
accelerator_type_3=InputValidator(
choices=["ml.eia1.medium", "ml.eia1.large", "ml.eia1.xlarge", ""],
input_type=str,
required=False,
description=
"The size of the Elastic Inference (EI) instance to use for the production variant.",
default="",
),
resource_encryption_key=InputValidator(
input_type=str,
required=False,
description=
"The AWS KMS key that Amazon SageMaker uses to encrypt data on the storage volume attached to the ML compute instance(s).",
default="",
),
endpoint_config_tags=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"An array of key-value pairs, to categorize AWS resources.",
default={},
),
endpoint_name=InputValidator(
input_type=str,
required=False,
description="The name of the endpoint.",
default="",
),
**vars(COMMON_INPUTS),
)
OUTPUTS = SageMakerDeployOutputs(endpoint_name=OutputValidator(
description="The created endpoint name."), )
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerDeployInputs,
SageMakerDeployOutputs)
@property
def inputs(self) -> SageMakerDeployInputs:
return self._inputs
@property
def outputs(self) -> SageMakerDeployOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerDeployOutputs:
return self._output_paths
class SageMakerRLEstimatorSpec(
SageMakerComponentSpec[SageMakerRLEstimatorInputs,
SageMakerRLEstimatorOutputs]):
INPUTS: SageMakerRLEstimatorInputs = SageMakerRLEstimatorInputs(
job_name=InputValidator(
input_type=str,
required=False,
description="Training job name.",
default="",
),
role=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) that Amazon SageMaker assumes to perform tasks on your behalf.",
),
image=InputValidator(
input_type=str,
required=False,
description=
"An ECR url. If specified, the estimator will use this image for training and hosting",
default=None,
),
entry_point=InputValidator(
input_type=str,
required=True,
description=
"Path (absolute or relative) to the Python source file which should be executed as the entry point to training.",
default="",
),
source_dir=InputValidator(
input_type=str,
required=False,
description=
"Path (S3 URI) to a directory with any other training source code dependencies aside from the entry point file.",
default="",
),
toolkit=InputValidator(
input_type=str,
choices=["coach", "ray", ""],
required=False,
description=
"RL toolkit you want to use for executing your model training code.",
default="",
),
toolkit_version=InputValidator(
input_type=str,
required=False,
description=
"RL toolkit version you want to be use for executing your model training code.",
default=None,
),
framework=InputValidator(
input_type=str,
choices=["tensorflow", "mxnet", "pytorch", ""],
required=False,
description=
"Framework (MXNet, TensorFlow or PyTorch) you want to be used as a toolkit backed for reinforcement learning training.",
default="",
),
metric_definitions=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The dictionary of name-regex pairs specify the metrics that the algorithm emits.",
default=[],
),
training_input_mode=InputValidator(
choices=["File", "Pipe"],
input_type=str,
description=
"The input mode that the algorithm supports. File or Pipe.",
default="File",
),
hyperparameters=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description="Hyperparameters that will be used for training.",
default={},
),
instance_type=InputValidator(
input_type=str,
required=False,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
instance_count=InputValidator(
input_type=int,
required=False,
description=
"The number of ML compute instances to use in the training job.",
default=1,
),
volume_size=InputValidator(
input_type=int,
required=True,
description=
"The size of the ML storage volume that you want to provision.",
default=30,
),
max_run=InputValidator(
input_type=int,
required=False,
description=
"Timeout in seconds for training (default: 24 * 60 * 60).",
default=24 * 60 * 60,
),
model_artifact_path=InputValidator(
input_type=str,
required=True,
description=
"Identifies the S3 path where you want Amazon SageMaker to store the model artifacts.",
),
vpc_security_group_ids=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description="The VPC security group IDs, in the form sg-xxxxxxxx.",
default=[],
),
vpc_subnets=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The ID of the subnets in the VPC to which you want to connect your hpo job.",
default=[],
),
network_isolation=InputValidator(
input_type=SpecInputParsers.str_to_bool,
description="Isolates the training container.",
default=False,
),
traffic_encryption=InputValidator(
input_type=SpecInputParsers.str_to_bool,
description=
"Encrypts all communications between ML compute instances in distributed training.",
default=False,
),
debug_hook_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"Configuration information for the debug hook parameters, collection configuration, and storage paths.",
default={},
),
debug_rule_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description="Configuration information for debugging rules.",
default=[],
),
**vars(COMMON_INPUTS),
**vars(SPOT_INSTANCE_INPUTS))
OUTPUTS = SageMakerRLEstimatorOutputs(
model_artifact_url=OutputValidator(
description="The model artifacts URL."),
job_name=OutputValidator(description="The training job name."),
training_image=OutputValidator(
description=
"The registry path of the Docker image that contains the training algorithm."
),
)
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerRLEstimatorInputs,
SageMakerRLEstimatorOutputs)
@property
def inputs(self) -> SageMakerRLEstimatorInputs:
return self._inputs
@property
def outputs(self) -> SageMakerRLEstimatorOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerRLEstimatorOutputs:
return self._output_paths
class SageMakerTrainingSpec(
SageMakerComponentSpec[SageMakerTrainingInputs, SageMakerTrainingOutputs]
):
INPUTS: SageMakerTrainingInputs = SageMakerTrainingInputs(
job_name=InputValidator(
input_type=str, description="The name of the training job.", default="",
),
role=InputValidator(
input_type=str,
required=True,
description="The Amazon Resource Name (ARN) that Amazon SageMaker assumes to perform tasks on your behalf.",
),
image=InputValidator(
input_type=str,
description="The registry path of the Docker image that contains the training algorithm.",
default="",
),
algorithm_name=InputValidator(
input_type=str,
description="The name of the resource algorithm to use for the training job. Do not specify a value for this if using training image.",
default="",
),
metric_definitions=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
description="The dictionary of name-regex pairs specify the metrics that the algorithm emits.",
default={},
),
training_input_mode=InputValidator(
choices=["File", "Pipe"],
input_type=str,
description="The input mode that the algorithm supports. File or Pipe.",
default="File",
),
hyperparameters=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
description="Dictionary of hyperparameters for the the algorithm.",
default={},
),
channels=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=True,
description="A list of dicts specifying the input channels. Must have at least one.",
),
instance_type=InputValidator(
input_type=str,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
instance_count=InputValidator(
required=True,
input_type=int,
description="The registry path of the Docker image that contains the training algorithm.",
default=1,
),
volume_size=InputValidator(
input_type=int,
required=True,
description="The size of the ML storage volume that you want to provision.",
default=30,
),
resource_encryption_key=InputValidator(
input_type=str,
description="The AWS KMS key that Amazon SageMaker uses to encrypt data on the storage volume attached to the ML compute instance(s).",
default="",
),
model_artifact_path=InputValidator(
input_type=str,
required=True,
description="Identifies the S3 path where you want Amazon SageMaker to store the model artifacts.",
),
output_encryption_key=InputValidator(
input_type=str,
description="The AWS KMS key that Amazon SageMaker uses to encrypt the model artifacts.",
default="",
),
vpc_security_group_ids=InputValidator(
input_type=str,
description="The VPC security group IDs, in the form sg-xxxxxxxx.",
),
vpc_subnets=InputValidator(
input_type=str,
description="The ID of the subnets in the VPC to which you want to connect your hpo job.",
),
network_isolation=InputValidator(
input_type=SpecInputParsers.str_to_bool,
description="Isolates the training container.",
default=True,
),
traffic_encryption=InputValidator(
input_type=SpecInputParsers.str_to_bool,
description="Encrypts all communications between ML compute instances in distributed training.",
default=False,
),
debug_hook_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
description="Configuration information for the debug hook parameters, collection configuration, and storage paths.",
default={},
),
debug_rule_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
description="Configuration information for debugging rules.",
default=[],
),
**vars(COMMON_INPUTS),
**vars(SPOT_INSTANCE_INPUTS)
)
OUTPUTS = SageMakerTrainingOutputs(
model_artifact_url=OutputValidator(description="The model artifacts URL."),
job_name=OutputValidator(description="The training job name."),
training_image=OutputValidator(
description="The registry path of the Docker image that contains the training algorithm."
),
)
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerTrainingInputs, SageMakerTrainingOutputs)
@property
def inputs(self) -> SageMakerTrainingInputs:
return self._inputs
@property
def outputs(self) -> SageMakerTrainingOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerTrainingOutputs:
return self._output_paths
class SageMakerProcessSpec(SageMakerComponentSpec[SageMakerProcessInputs,
SageMakerProcessOutputs]):
INPUTS: SageMakerProcessInputs = SageMakerProcessInputs(
job_name=InputValidator(
input_type=str,
required=False,
description="The name of the processing job.",
default="",
),
role=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) that Amazon SageMaker assumes to perform tasks on your behalf.",
),
image=InputValidator(
input_type=str,
required=True,
description=
"The registry path of the Docker image that contains the processing container.",
default="",
),
instance_type=InputValidator(
required=True,
input_type=str,
description="The ML compute instance type.",
default="ml.m4.xlarge",
),
instance_count=InputValidator(
required=True,
input_type=int,
description=
"The number of ML compute instances to use in each processing job.",
default=1,
),
volume_size=InputValidator(
input_type=int,
required=False,
description=
"The size of the ML storage volume that you want to provision.",
default=30,
),
resource_encryption_key=InputValidator(
input_type=str,
required=False,
description=
"The AWS KMS key that Amazon SageMaker uses to encrypt data on the storage volume attached to the ML compute instance(s).",
default="",
),
output_encryption_key=InputValidator(
input_type=str,
required=False,
description=
"The AWS KMS key that Amazon SageMaker uses to encrypt the processing artifacts.",
default="",
),
max_run_time=InputValidator(
input_type=int,
required=False,
description=
"The maximum run time in seconds for the processing job.",
default=86400,
),
environment=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"The dictionary of the environment variables to set in the Docker container. Up to 16 key-value entries in the map.",
default={},
),
container_entrypoint=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The entrypoint for the processing job. This is in the form of a list of strings that make a command.",
default=[],
),
container_arguments=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"A list of string arguments to be passed to a processing job.",
default=[],
),
input_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"Parameters that specify Amazon S3 inputs for a processing job.",
default=[],
),
output_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=True,
description=
"Parameters that specify Amazon S3 outputs for a processing job.",
default=[],
),
vpc_security_group_ids=InputValidator(
input_type=str,
required=False,
description="The VPC security group IDs, in the form sg-xxxxxxxx.",
),
vpc_subnets=InputValidator(
input_type=str,
required=False,
description=
"The ID of the subnets in the VPC to which you want to connect your hpo job.",
),
network_isolation=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="Isolates the processing container.",
default=True,
),
traffic_encryption=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description=
"Encrypts all communications between ML compute instances in distributed training.",
default=False,
),
**vars(COMMON_INPUTS),
)
OUTPUTS = SageMakerProcessOutputs(
job_name=OutputValidator(description="Processing job name."),
output_artifacts=OutputValidator(
description="A dictionary containing the output S3 artifacts."),
)
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerProcessInputs,
SageMakerProcessOutputs)
@property
def inputs(self) -> SageMakerProcessInputs:
return self._inputs
@property
def outputs(self) -> SageMakerProcessOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerProcessOutputs:
return self._output_paths
class SageMakerCreateModelSpec(
SageMakerComponentSpec[SageMakerCreateModelInputs,
SageMakerCreateModelOutputs]):
INPUTS: SageMakerCreateModelInputs = SageMakerCreateModelInputs(
model_name=InputValidator(input_type=str,
required=True,
description="The name of the new model."),
role=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) that Amazon SageMaker assumes to perform tasks on your behalf.",
),
container_host_name=InputValidator(
input_type=str,
required=False,
description=
"When a ContainerDefinition is part of an inference pipeline, this value uniquely identifies the container for the purposes of logging and metrics.",
default="",
),
image=InputValidator(
input_type=str,
required=False,
description=
"The Amazon EC2 Container Registry (Amazon ECR) path where inference code is stored.",
default="",
),
model_artifact_url=InputValidator(
input_type=str,
required=False,
description=
"S3 path where Amazon SageMaker to store the model artifacts.",
default="",
),
environment=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"The dictionary of the environment variables to set in the Docker container. Up to 16 key-value entries in the map.",
default={},
),
model_package=InputValidator(
input_type=str,
required=False,
description=
"The name or Amazon Resource Name (ARN) of the model package to use to create the model.",
default="",
),
secondary_containers=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"A list of dicts that specifies the additional containers in the inference pipeline.",
default=[],
),
vpc_security_group_ids=InputValidator(
input_type=str,
required=False,
description="The VPC security group IDs, in the form sg-xxxxxxxx.",
default="",
),
vpc_subnets=InputValidator(
input_type=str,
required=False,
description=
"The ID of the subnets in the VPC to which you want to connect your hpo job.",
default="",
),
network_isolation=InputValidator(
input_type=SpecInputParsers.str_to_bool,
required=False,
description="Isolates the training container.",
default=True,
),
**vars(COMMON_INPUTS),
)
OUTPUTS = SageMakerCreateModelOutputs(model_name=OutputValidator(
description="The name of the model created by SageMaker."), )
def __init__(self, arguments: List[str]):
super().__init__(arguments, SageMakerCreateModelInputs,
SageMakerCreateModelOutputs)
@property
def inputs(self) -> SageMakerCreateModelInputs:
return self._inputs
@property
def outputs(self) -> SageMakerCreateModelOutputs:
return self._outputs
@property
def output_paths(self) -> SageMakerCreateModelOutputs:
return self._output_paths
class RoboMakerSimulationJobSpec(
SageMakerComponentSpec[RoboMakerSimulationJobInputs,
RoboMakerSimulationJobOutputs]):
INPUTS: RoboMakerSimulationJobInputs = RoboMakerSimulationJobInputs(
role=InputValidator(
input_type=str,
required=True,
description=
"The Amazon Resource Name (ARN) that Amazon RoboMaker assumes to perform tasks on your behalf.",
),
output_bucket=InputValidator(
input_type=str,
required=True,
description="The bucket to place outputs from the simulation job.",
default="",
),
output_path=InputValidator(
input_type=str,
required=True,
description=
"The S3 key where outputs from the simulation job are placed.",
default="",
),
max_run=InputValidator(
input_type=int,
required=True,
description=
"Timeout in seconds for simulation job (default: 8 * 60 * 60).",
default=8 * 60 * 60,
),
failure_behavior=InputValidator(
input_type=str,
required=False,
description=
"The failure behavior the simulation job (Continue|Fail).",
default="Fail",
),
sim_app_arn=InputValidator(
input_type=str,
required=False,
description="The application ARN for the simulation application.",
default="",
),
sim_app_version=InputValidator(
input_type=str,
required=False,
description=
"The application version for the simulation application.",
default="",
),
sim_app_launch_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description=
"The launch configuration for the simulation application.",
default={},
),
sim_app_world_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description="A list of world configurations.",
default=[],
),
robot_app_arn=InputValidator(
input_type=str,
required=False,
description="The application ARN for the robot application.",
default="",
),
robot_app_version=InputValidator(
input_type=str,
required=False,
description="The application version for the robot application.",
default="",
),
robot_app_launch_config=InputValidator(
input_type=SpecInputParsers.yaml_or_json_dict,
required=False,
description="The launch configuration for the robot application.",
default={},
),
data_sources=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"Specify data sources to mount read-only files from S3 into your simulation.",
default=[],
),
vpc_security_group_ids=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description="The VPC security group IDs, in the form sg-xxxxxxxx.",
default=[],
),
vpc_subnets=InputValidator(
input_type=SpecInputParsers.yaml_or_json_list,
required=False,
description=
"The ID of the subnets in the VPC to which you want to connect your simulation job.",
default=[],
),
use_public_ip=InputValidator(
input_type=bool,
description=
"A boolean indicating whether to assign a public IP address.",
default=False,
),
sim_unit_limit=InputValidator(
input_type=int,
required=False,
description="The simulation unit limit.",
default=15,
),
record_ros_topics=InputValidator(
input_type=bool,
description=
"A boolean indicating whether to record all ROS topics. Used for logging.",
default=False,
),
**vars(COMMON_INPUTS),
)
OUTPUTS = RoboMakerSimulationJobOutputs(
arn=OutputValidator(
description="The Amazon Resource Name (ARN) of the simulation job."
),
output_artifacts=OutputValidator(
description="The simulation job artifacts URL."),
job_id=OutputValidator(description="The simulation job id."),
)
def __init__(self, arguments: List[str]):
super().__init__(
arguments,
RoboMakerSimulationJobInputs,
RoboMakerSimulationJobOutputs,
)
@property
def inputs(self) -> RoboMakerSimulationJobInputs:
return self._inputs
@property
def outputs(self) -> RoboMakerSimulationJobOutputs:
return self._outputs
@property
def output_paths(self) -> RoboMakerSimulationJobOutputs:
return self._output_paths