Amazonka.SageMaker.Types.HyperbandStrategyConfig

Documentation

data HyperbandStrategyConfig

The configuration for Hyperband, a multi-fidelity based hyperparameter tuning strategy. Hyperband uses the final and intermediate results of a training job to dynamically allocate resources to utilized hyperparameter configurations while automatically stopping under-performing configurations. This parameter should be provided only if Hyperband is selected as the StrategyConfig under the HyperParameterTuningJobConfig API.

See: newHyperbandStrategyConfig smart constructor.

Constructors

HyperbandStrategyConfig'

Fields

• maxResource :: Maybe Natural

  The maximum number of resources (such as epochs) that can be used by a training job launched by a hyperparameter tuning job. Once a job reaches the MaxResource value, it is stopped. If a value for MaxResource is not provided, and Hyperband is selected as the hyperparameter tuning strategy, HyperbandTrainingJ attempts to infer MaxResource from the following keys (if present) in StaticsHyperParameters:

  • epochs

  • numepochs

  • n-epochs

  • n_epochs

  • num_epochs

  If HyperbandStrategyConfig is unable to infer a value for MaxResource, it generates a validation error. The maximum value is 20,000 epochs. All metrics that correspond to an objective metric are used to derive early stopping decisions. For distributive training jobs, ensure that duplicate metrics are not printed in the logs across the individual nodes in a training job. If multiple nodes are publishing duplicate or incorrect metrics, training jobs may make an incorrect stopping decision and stop the job prematurely.

• minResource :: Maybe Natural

  The minimum number of resources (such as epochs) that can be used by a training job launched by a hyperparameter tuning job. If the value for MinResource has not been reached, the training job will not be stopped by Hyperband.

Instances

FromJSON HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

parseJSON :: Value -> Parser HyperbandStrategyConfig

parseJSONList :: Value -> Parser [HyperbandStrategyConfig]

ToJSON HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

toJSON :: HyperbandStrategyConfig -> Value

toEncoding :: HyperbandStrategyConfig -> Encoding

toJSONList :: [HyperbandStrategyConfig] -> Value

toEncodingList :: [HyperbandStrategyConfig] -> Encoding

Generic HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Associated Types

type Rep HyperbandStrategyConfig :: Type -> Type

Methods

from :: HyperbandStrategyConfig -> Rep HyperbandStrategyConfig x

to :: Rep HyperbandStrategyConfig x -> HyperbandStrategyConfig

Read HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

readsPrec :: Int -> ReadS HyperbandStrategyConfig

readList :: ReadS [HyperbandStrategyConfig]

readPrec :: ReadPrec HyperbandStrategyConfig

readListPrec :: ReadPrec [HyperbandStrategyConfig]

Show HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

showsPrec :: Int -> HyperbandStrategyConfig -> ShowS

show :: HyperbandStrategyConfig -> String

showList :: [HyperbandStrategyConfig] -> ShowS

NFData HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

rnf :: HyperbandStrategyConfig -> ()

Eq HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

(==) :: HyperbandStrategyConfig -> HyperbandStrategyConfig -> Bool

(/=) :: HyperbandStrategyConfig -> HyperbandStrategyConfig -> Bool

Hashable HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

Methods

hashWithSalt :: Int -> HyperbandStrategyConfig -> Int

hash :: HyperbandStrategyConfig -> Int

type Rep HyperbandStrategyConfig

Defined in Amazonka.SageMaker.Types.HyperbandStrategyConfig

type Rep HyperbandStrategyConfig = D1 ('MetaData "HyperbandStrategyConfig" "Amazonka.SageMaker.Types.HyperbandStrategyConfig" "amazonka-sagemaker-2.0-9SyrKZ4KqhsL1qX9u3ILA3" 'False) (C1 ('MetaCons "HyperbandStrategyConfig'" 'PrefixI 'True) (S1 ('MetaSel ('Just "maxResource") 'NoSourceUnpackedness 'NoSourceStrictness 'DecidedStrict) (Rec0 (Maybe Natural)) :*: S1 ('MetaSel ('Just "minResource") 'NoSourceUnpackedness 'NoSourceStrictness 'DecidedStrict) (Rec0 (Maybe Natural))))

newHyperbandStrategyConfig :: HyperbandStrategyConfig

Create a value of HyperbandStrategyConfig with all optional fields omitted.

Use generic-lens or optics to modify other optional fields.

The following record fields are available, with the corresponding lenses provided for backwards compatibility:

$sel:maxResource:HyperbandStrategyConfig', hyperbandStrategyConfig_maxResource - The maximum number of resources (such as epochs) that can be used by a training job launched by a hyperparameter tuning job. Once a job reaches the MaxResource value, it is stopped. If a value for MaxResource is not provided, and Hyperband is selected as the hyperparameter tuning strategy, HyperbandTrainingJ attempts to infer MaxResource from the following keys (if present) in StaticsHyperParameters:

• epochs

• numepochs

• n-epochs

• n_epochs

• num_epochs

If HyperbandStrategyConfig is unable to infer a value for MaxResource, it generates a validation error. The maximum value is 20,000 epochs. All metrics that correspond to an objective metric are used to derive early stopping decisions. For distributive training jobs, ensure that duplicate metrics are not printed in the logs across the individual nodes in a training job. If multiple nodes are publishing duplicate or incorrect metrics, training jobs may make an incorrect stopping decision and stop the job prematurely.

$sel:minResource:HyperbandStrategyConfig', hyperbandStrategyConfig_minResource - The minimum number of resources (such as epochs) that can be used by a training job launched by a hyperparameter tuning job. If the value for MinResource has not been reached, the training job will not be stopped by Hyperband.

hyperbandStrategyConfig_maxResource :: Lens' HyperbandStrategyConfig (Maybe Natural)

The maximum number of resources (such as epochs) that can be used by a training job launched by a hyperparameter tuning job. Once a job reaches the MaxResource value, it is stopped. If a value for MaxResource is not provided, and Hyperband is selected as the hyperparameter tuning strategy, HyperbandTrainingJ attempts to infer MaxResource from the following keys (if present) in StaticsHyperParameters:

• epochs

• numepochs

• n-epochs

• n_epochs

• num_epochs

If HyperbandStrategyConfig is unable to infer a value for MaxResource, it generates a validation error. The maximum value is 20,000 epochs. All metrics that correspond to an objective metric are used to derive early stopping decisions. For distributive training jobs, ensure that duplicate metrics are not printed in the logs across the individual nodes in a training job. If multiple nodes are publishing duplicate or incorrect metrics, training jobs may make an incorrect stopping decision and stop the job prematurely.

hyperbandStrategyConfig_minResource :: Lens' HyperbandStrategyConfig (Maybe Natural)

The minimum number of resources (such as epochs) that can be used by a training job launched by a hyperparameter tuning job. If the value for MinResource has not been reached, the training job will not be stopped by Hyperband.