aws-sdk-sagemaker 1.138.0 → 1.140.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 289e07fa18ad373afbd8939e3fc8b3bf59bcb129e3f605a18604607e4389672a
4
- data.tar.gz: 12d3c4374927f455fe31c21e33ccb43e52645fc54fa911cad629dd20bc1d4c4d
3
+ metadata.gz: 183a3d0710845a318335195f1df77b55daec4ac09cce10e3f8d78aa502900df8
4
+ data.tar.gz: c52b1b8802e3bb268c7461a679e8a40621c2a987889458371e59595708deadbc
5
5
  SHA512:
6
- metadata.gz: 29108cd59a45026cf8c99f5917bf1ae9546dc8777275fc2273fcab251aad1652b055fe259d708f82ef60fd89f7d2c859496541ae279dfa6c6d23efaddf3f663a
7
- data.tar.gz: 6ee2ec14772d5fd21e1b4a0551a16c6b92d13bf4d715677fb7d6e0804750e695828082d6fd83ad83f46c42c11eee2a346a85b660b9d5080639f291ffe88fbb85
6
+ metadata.gz: 272cc69827d58efd70a692bd40bfbae65b89fd7e21db5e817ad9fa26785b2001d616a9797c720a7ab88ae5e32b07bd9237edd5dfc8800227d6d4d8385f98ef73
7
+ data.tar.gz: 87c6ec30a6c56ae2f46ad9ae99f26aded31705b658d707917940ba61a5b90c9423ebf890cf7bf3084fc1e48e763145c911b7094de28cbf259fb8b994a1781063
data/CHANGELOG.md CHANGED
@@ -1,6 +1,16 @@
1
1
  Unreleased Changes
2
2
  ------------------
3
3
 
4
+ 1.140.0 (2022-09-15)
5
+ ------------------
6
+
7
+ * Feature - Amazon SageMaker Automatic Model Tuning now supports specifying Hyperband strategy for tuning jobs, which uses a multi-fidelity based tuning strategy to stop underperforming hyperparameter configurations early.
8
+
9
+ 1.139.0 (2022-09-08)
10
+ ------------------
11
+
12
+ * Feature - This release adds Mode to AutoMLJobConfig.
13
+
4
14
  1.138.0 (2022-09-07)
5
15
  ------------------
6
16
 
data/VERSION CHANGED
@@ -1 +1 @@
1
- 1.138.0
1
+ 1.140.0
@@ -940,8 +940,7 @@ module Aws::SageMaker
940
940
  req.send_request(options)
941
941
  end
942
942
 
943
- # Creates a running app for the specified UserProfile. Supported apps
944
- # are `JupyterServer` and `KernelGateway`. This operation is
943
+ # Creates a running app for the specified UserProfile. This operation is
945
944
  # automatically invoked by Amazon SageMaker Studio upon access to the
946
945
  # associated Domain, and when new kernel configurations are selected by
947
946
  # the user. A user may have multiple Apps active simultaneously.
@@ -953,8 +952,7 @@ module Aws::SageMaker
953
952
  # The user profile name.
954
953
  #
955
954
  # @option params [required, String] :app_type
956
- # The type of app. Supported apps are `JupyterServer` and
957
- # `KernelGateway`. `TensorBoard` is not supported.
955
+ # The type of app.
958
956
  #
959
957
  # @option params [required, String] :app_name
960
958
  # The name of the app.
@@ -1260,6 +1258,7 @@ module Aws::SageMaker
1260
1258
  # candidate_generation_config: {
1261
1259
  # feature_specification_s3_uri: "S3Uri",
1262
1260
  # },
1261
+ # mode: "AUTO", # accepts AUTO, ENSEMBLING, HYPERPARAMETER_TUNING
1263
1262
  # },
1264
1263
  # role_arn: "RoleArn", # required
1265
1264
  # generate_candidate_definitions_only: false,
@@ -3014,7 +3013,13 @@ module Aws::SageMaker
3014
3013
  # resp = client.create_hyper_parameter_tuning_job({
3015
3014
  # hyper_parameter_tuning_job_name: "HyperParameterTuningJobName", # required
3016
3015
  # hyper_parameter_tuning_job_config: { # required
3017
- # strategy: "Bayesian", # required, accepts Bayesian, Random
3016
+ # strategy: "Bayesian", # required, accepts Bayesian, Random, Hyperband
3017
+ # strategy_config: {
3018
+ # hyperband_strategy_config: {
3019
+ # min_resource: 1,
3020
+ # max_resource: 1,
3021
+ # },
3022
+ # },
3018
3023
  # hyper_parameter_tuning_job_objective: {
3019
3024
  # type: "Maximize", # required, accepts Maximize, Minimize
3020
3025
  # metric_name: "MetricName", # required
@@ -5834,6 +5839,12 @@ module Aws::SageMaker
5834
5839
  # a list of hyperparameters for each training algorithm provided by
5835
5840
  # SageMaker, see [Algorithms][1].
5836
5841
  #
5842
+ # You must not include any security-sensitive information, such as
5843
+ # account access IDs, secrets, and tokens, in the dictionary for
5844
+ # configuring hyperparameters. SageMaker rejects the training job
5845
+ # request and returns an exception error for detected credentials, if
5846
+ # such user input is found.
5847
+ #
5837
5848
  # * `InputDataConfig` - Describes the training dataset and the Amazon
5838
5849
  # S3, EFS, or FSx location where it is stored.
5839
5850
  #
@@ -5886,6 +5897,12 @@ module Aws::SageMaker
5886
5897
  # is a key-value pair. Each key and value is limited to 256 characters,
5887
5898
  # as specified by the `Length Constraint`.
5888
5899
  #
5900
+ # You must not include any security-sensitive information, such as
5901
+ # account access IDs, secrets, and tokens, in the dictionary for
5902
+ # configuring hyperparameters. SageMaker rejects the training job
5903
+ # request and returns an exception error for detected credentials, if
5904
+ # such user input is found.
5905
+ #
5889
5906
  #
5890
5907
  #
5891
5908
  # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html
@@ -8571,6 +8588,7 @@ module Aws::SageMaker
8571
8588
  # resp.auto_ml_job_config.security_config.vpc_config.subnets[0] #=> String
8572
8589
  # resp.auto_ml_job_config.data_split_config.validation_fraction #=> Float
8573
8590
  # resp.auto_ml_job_config.candidate_generation_config.feature_specification_s3_uri #=> String
8591
+ # resp.auto_ml_job_config.mode #=> String, one of "AUTO", "ENSEMBLING", "HYPERPARAMETER_TUNING"
8574
8592
  # resp.creation_time #=> Time
8575
8593
  # resp.end_time #=> Time
8576
8594
  # resp.last_modified_time #=> Time
@@ -9708,7 +9726,9 @@ module Aws::SageMaker
9708
9726
  #
9709
9727
  # resp.hyper_parameter_tuning_job_name #=> String
9710
9728
  # resp.hyper_parameter_tuning_job_arn #=> String
9711
- # resp.hyper_parameter_tuning_job_config.strategy #=> String, one of "Bayesian", "Random"
9729
+ # resp.hyper_parameter_tuning_job_config.strategy #=> String, one of "Bayesian", "Random", "Hyperband"
9730
+ # resp.hyper_parameter_tuning_job_config.strategy_config.hyperband_strategy_config.min_resource #=> Integer
9731
+ # resp.hyper_parameter_tuning_job_config.strategy_config.hyperband_strategy_config.max_resource #=> Integer
9712
9732
  # resp.hyper_parameter_tuning_job_config.hyper_parameter_tuning_job_objective.type #=> String, one of "Maximize", "Minimize"
9713
9733
  # resp.hyper_parameter_tuning_job_config.hyper_parameter_tuning_job_objective.metric_name #=> String
9714
9734
  # resp.hyper_parameter_tuning_job_config.resource_limits.max_number_of_training_jobs #=> Integer
@@ -14064,7 +14084,7 @@ module Aws::SageMaker
14064
14084
  # resp.hyper_parameter_tuning_job_summaries[0].hyper_parameter_tuning_job_name #=> String
14065
14085
  # resp.hyper_parameter_tuning_job_summaries[0].hyper_parameter_tuning_job_arn #=> String
14066
14086
  # resp.hyper_parameter_tuning_job_summaries[0].hyper_parameter_tuning_job_status #=> String, one of "Completed", "InProgress", "Failed", "Stopped", "Stopping"
14067
- # resp.hyper_parameter_tuning_job_summaries[0].strategy #=> String, one of "Bayesian", "Random"
14087
+ # resp.hyper_parameter_tuning_job_summaries[0].strategy #=> String, one of "Bayesian", "Random", "Hyperband"
14068
14088
  # resp.hyper_parameter_tuning_job_summaries[0].creation_time #=> Time
14069
14089
  # resp.hyper_parameter_tuning_job_summaries[0].hyper_parameter_tuning_end_time #=> Time
14070
14090
  # resp.hyper_parameter_tuning_job_summaries[0].last_modified_time #=> Time
@@ -17933,7 +17953,9 @@ module Aws::SageMaker
17933
17953
  # resp.results[0].feature_metadata.parameters[0].value #=> String
17934
17954
  # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_name #=> String
17935
17955
  # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_arn #=> String
17936
- # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.strategy #=> String, one of "Bayesian", "Random"
17956
+ # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.strategy #=> String, one of "Bayesian", "Random", "Hyperband"
17957
+ # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.strategy_config.hyperband_strategy_config.min_resource #=> Integer
17958
+ # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.strategy_config.hyperband_strategy_config.max_resource #=> Integer
17937
17959
  # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.hyper_parameter_tuning_job_objective.type #=> String, one of "Maximize", "Minimize"
17938
17960
  # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.hyper_parameter_tuning_job_objective.metric_name #=> String
17939
17961
  # resp.results[0].hyper_parameter_tuning_job.hyper_parameter_tuning_job_config.resource_limits.max_number_of_training_jobs #=> Integer
@@ -20582,7 +20604,7 @@ module Aws::SageMaker
20582
20604
  params: params,
20583
20605
  config: config)
20584
20606
  context[:gem_name] = 'aws-sdk-sagemaker'
20585
- context[:gem_version] = '1.138.0'
20607
+ context[:gem_version] = '1.140.0'
20586
20608
  Seahorse::Client::Request.new(handlers, context)
20587
20609
  end
20588
20610
 
@@ -121,6 +121,7 @@ module Aws::SageMaker
121
121
  AutoMLMaxResults = Shapes::IntegerShape.new(name: 'AutoMLMaxResults')
122
122
  AutoMLMetricEnum = Shapes::StringShape.new(name: 'AutoMLMetricEnum')
123
123
  AutoMLMetricExtendedEnum = Shapes::StringShape.new(name: 'AutoMLMetricExtendedEnum')
124
+ AutoMLMode = Shapes::StringShape.new(name: 'AutoMLMode')
124
125
  AutoMLNameContains = Shapes::StringShape.new(name: 'AutoMLNameContains')
125
126
  AutoMLOutputDataConfig = Shapes::StructureShape.new(name: 'AutoMLOutputDataConfig')
126
127
  AutoMLPartialFailureReason = Shapes::StructureShape.new(name: 'AutoMLPartialFailureReason')
@@ -753,6 +754,7 @@ module Aws::SageMaker
753
754
  HyperParameterTuningJobSearchEntity = Shapes::StructureShape.new(name: 'HyperParameterTuningJobSearchEntity')
754
755
  HyperParameterTuningJobSortByOptions = Shapes::StringShape.new(name: 'HyperParameterTuningJobSortByOptions')
755
756
  HyperParameterTuningJobStatus = Shapes::StringShape.new(name: 'HyperParameterTuningJobStatus')
757
+ HyperParameterTuningJobStrategyConfig = Shapes::StructureShape.new(name: 'HyperParameterTuningJobStrategyConfig')
756
758
  HyperParameterTuningJobStrategyType = Shapes::StringShape.new(name: 'HyperParameterTuningJobStrategyType')
757
759
  HyperParameterTuningJobSummaries = Shapes::ListShape.new(name: 'HyperParameterTuningJobSummaries')
758
760
  HyperParameterTuningJobSummary = Shapes::StructureShape.new(name: 'HyperParameterTuningJobSummary')
@@ -761,6 +763,9 @@ module Aws::SageMaker
761
763
  HyperParameterTuningResourceConfig = Shapes::StructureShape.new(name: 'HyperParameterTuningResourceConfig')
762
764
  HyperParameterValue = Shapes::StringShape.new(name: 'HyperParameterValue')
763
765
  HyperParameters = Shapes::MapShape.new(name: 'HyperParameters')
766
+ HyperbandStrategyConfig = Shapes::StructureShape.new(name: 'HyperbandStrategyConfig')
767
+ HyperbandStrategyMaxResource = Shapes::IntegerShape.new(name: 'HyperbandStrategyMaxResource')
768
+ HyperbandStrategyMinResource = Shapes::IntegerShape.new(name: 'HyperbandStrategyMinResource')
764
769
  IdempotencyToken = Shapes::StringShape.new(name: 'IdempotencyToken')
765
770
  Image = Shapes::StructureShape.new(name: 'Image')
766
771
  ImageArn = Shapes::StringShape.new(name: 'ImageArn')
@@ -1952,6 +1957,7 @@ module Aws::SageMaker
1952
1957
  AutoMLJobConfig.add_member(:security_config, Shapes::ShapeRef.new(shape: AutoMLSecurityConfig, location_name: "SecurityConfig"))
1953
1958
  AutoMLJobConfig.add_member(:data_split_config, Shapes::ShapeRef.new(shape: AutoMLDataSplitConfig, location_name: "DataSplitConfig"))
1954
1959
  AutoMLJobConfig.add_member(:candidate_generation_config, Shapes::ShapeRef.new(shape: AutoMLCandidateGenerationConfig, location_name: "CandidateGenerationConfig"))
1960
+ AutoMLJobConfig.add_member(:mode, Shapes::ShapeRef.new(shape: AutoMLMode, location_name: "Mode"))
1955
1961
  AutoMLJobConfig.struct_class = Types::AutoMLJobConfig
1956
1962
 
1957
1963
  AutoMLJobObjective.add_member(:metric_name, Shapes::ShapeRef.new(shape: AutoMLMetricEnum, required: true, location_name: "MetricName"))
@@ -4519,6 +4525,7 @@ module Aws::SageMaker
4519
4525
  HyperParameterTuningInstanceConfigs.member = Shapes::ShapeRef.new(shape: HyperParameterTuningInstanceConfig)
4520
4526
 
4521
4527
  HyperParameterTuningJobConfig.add_member(:strategy, Shapes::ShapeRef.new(shape: HyperParameterTuningJobStrategyType, required: true, location_name: "Strategy"))
4528
+ HyperParameterTuningJobConfig.add_member(:strategy_config, Shapes::ShapeRef.new(shape: HyperParameterTuningJobStrategyConfig, location_name: "StrategyConfig"))
4522
4529
  HyperParameterTuningJobConfig.add_member(:hyper_parameter_tuning_job_objective, Shapes::ShapeRef.new(shape: HyperParameterTuningJobObjective, location_name: "HyperParameterTuningJobObjective"))
4523
4530
  HyperParameterTuningJobConfig.add_member(:resource_limits, Shapes::ShapeRef.new(shape: ResourceLimits, required: true, location_name: "ResourceLimits"))
4524
4531
  HyperParameterTuningJobConfig.add_member(:parameter_ranges, Shapes::ShapeRef.new(shape: ParameterRanges, location_name: "ParameterRanges"))
@@ -4550,6 +4557,9 @@ module Aws::SageMaker
4550
4557
  HyperParameterTuningJobSearchEntity.add_member(:tags, Shapes::ShapeRef.new(shape: TagList, location_name: "Tags"))
4551
4558
  HyperParameterTuningJobSearchEntity.struct_class = Types::HyperParameterTuningJobSearchEntity
4552
4559
 
4560
+ HyperParameterTuningJobStrategyConfig.add_member(:hyperband_strategy_config, Shapes::ShapeRef.new(shape: HyperbandStrategyConfig, location_name: "HyperbandStrategyConfig"))
4561
+ HyperParameterTuningJobStrategyConfig.struct_class = Types::HyperParameterTuningJobStrategyConfig
4562
+
4553
4563
  HyperParameterTuningJobSummaries.member = Shapes::ShapeRef.new(shape: HyperParameterTuningJobSummary)
4554
4564
 
4555
4565
  HyperParameterTuningJobSummary.add_member(:hyper_parameter_tuning_job_name, Shapes::ShapeRef.new(shape: HyperParameterTuningJobName, required: true, location_name: "HyperParameterTuningJobName"))
@@ -4579,6 +4589,10 @@ module Aws::SageMaker
4579
4589
  HyperParameters.key = Shapes::ShapeRef.new(shape: HyperParameterKey)
4580
4590
  HyperParameters.value = Shapes::ShapeRef.new(shape: HyperParameterValue)
4581
4591
 
4592
+ HyperbandStrategyConfig.add_member(:min_resource, Shapes::ShapeRef.new(shape: HyperbandStrategyMinResource, location_name: "MinResource"))
4593
+ HyperbandStrategyConfig.add_member(:max_resource, Shapes::ShapeRef.new(shape: HyperbandStrategyMaxResource, location_name: "MaxResource"))
4594
+ HyperbandStrategyConfig.struct_class = Types::HyperbandStrategyConfig
4595
+
4582
4596
  Image.add_member(:creation_time, Shapes::ShapeRef.new(shape: Timestamp, required: true, location_name: "CreationTime"))
4583
4597
  Image.add_member(:description, Shapes::ShapeRef.new(shape: ImageDescription, location_name: "Description"))
4584
4598
  Image.add_member(:display_name, Shapes::ShapeRef.new(shape: ImageDisplayName, location_name: "DisplayName"))
@@ -2338,6 +2338,7 @@ module Aws::SageMaker
2338
2338
  # candidate_generation_config: {
2339
2339
  # feature_specification_s3_uri: "S3Uri",
2340
2340
  # },
2341
+ # mode: "AUTO", # accepts AUTO, ENSEMBLING, HYPERPARAMETER_TUNING
2341
2342
  # }
2342
2343
  #
2343
2344
  # @!attribute [rw] completion_criteria
@@ -2361,13 +2362,43 @@ module Aws::SageMaker
2361
2362
  # (optional).
2362
2363
  # @return [Types::AutoMLCandidateGenerationConfig]
2363
2364
  #
2365
+ # @!attribute [rw] mode
2366
+ # The method that Autopilot uses to train the data. You can either
2367
+ # specify the mode manually or let Autopilot choose for you based on
2368
+ # the dataset size by selecting `AUTO`. In `AUTO` mode, Autopilot
2369
+ # chooses `ENSEMBLING` for datasets smaller than 100 MB, and
2370
+ # `HYPERPARAMETER_TUNING` for larger ones.
2371
+ #
2372
+ # The `ENSEMBLING` mode uses a multi-stack ensemble model to predict
2373
+ # classification and regression tasks directly from your dataset. This
2374
+ # machine learning mode combines several base models to produce an
2375
+ # optimal predictive model. It then uses a stacking ensemble method to
2376
+ # combine predictions from contributing members. A multi-stack
2377
+ # ensemble model can provide better performance over a single model by
2378
+ # combining the predictive capabilities of multiple models. See
2379
+ # [Autopilot algorithm support][1] for a list of algorithms supported
2380
+ # by `ENSEMBLING` mode.
2381
+ #
2382
+ # The `HYPERPARAMETER_TUNING` (HPO) mode uses the best hyperparameters
2383
+ # to train the best version of a model. HPO will automatically select
2384
+ # an algorithm for the type of problem you want to solve. Then HPO
2385
+ # finds the best hyperparameters according to your objective metric.
2386
+ # See [Autopilot algorithm support][1] for a list of algorithms
2387
+ # supported by `HYPERPARAMETER_TUNING` mode.
2388
+ #
2389
+ #
2390
+ #
2391
+ # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-model-support-validation.html#autopilot-algorithm-suppprt
2392
+ # @return [String]
2393
+ #
2364
2394
  # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/AutoMLJobConfig AWS API Documentation
2365
2395
  #
2366
2396
  class AutoMLJobConfig < Struct.new(
2367
2397
  :completion_criteria,
2368
2398
  :security_config,
2369
2399
  :data_split_config,
2370
- :candidate_generation_config)
2400
+ :candidate_generation_config,
2401
+ :mode)
2371
2402
  SENSITIVE = []
2372
2403
  include Aws::Structure
2373
2404
  end
@@ -4409,8 +4440,7 @@ module Aws::SageMaker
4409
4440
  # @return [String]
4410
4441
  #
4411
4442
  # @!attribute [rw] app_type
4412
- # The type of app. Supported apps are `JupyterServer` and
4413
- # `KernelGateway`. `TensorBoard` is not supported.
4443
+ # The type of app.
4414
4444
  # @return [String]
4415
4445
  #
4416
4446
  # @!attribute [rw] app_name
@@ -4591,6 +4621,7 @@ module Aws::SageMaker
4591
4621
  # candidate_generation_config: {
4592
4622
  # feature_specification_s3_uri: "S3Uri",
4593
4623
  # },
4624
+ # mode: "AUTO", # accepts AUTO, ENSEMBLING, HYPERPARAMETER_TUNING
4594
4625
  # },
4595
4626
  # role_arn: "RoleArn", # required
4596
4627
  # generate_candidate_definitions_only: false,
@@ -6261,7 +6292,13 @@ module Aws::SageMaker
6261
6292
  # {
6262
6293
  # hyper_parameter_tuning_job_name: "HyperParameterTuningJobName", # required
6263
6294
  # hyper_parameter_tuning_job_config: { # required
6264
- # strategy: "Bayesian", # required, accepts Bayesian, Random
6295
+ # strategy: "Bayesian", # required, accepts Bayesian, Random, Hyperband
6296
+ # strategy_config: {
6297
+ # hyperband_strategy_config: {
6298
+ # min_resource: 1,
6299
+ # max_resource: 1,
6300
+ # },
6301
+ # },
6265
6302
  # hyper_parameter_tuning_job_objective: {
6266
6303
  # type: "Maximize", # required, accepts Maximize, Minimize
6267
6304
  # metric_name: "MetricName", # required
@@ -9383,6 +9420,12 @@ module Aws::SageMaker
9383
9420
  # hyperparameter is a key-value pair. Each key and value is limited to
9384
9421
  # 256 characters, as specified by the `Length Constraint`.
9385
9422
  #
9423
+ # You must not include any security-sensitive information, such as
9424
+ # account access IDs, secrets, and tokens, in the dictionary for
9425
+ # configuring hyperparameters. SageMaker rejects the training job
9426
+ # request and returns an exception error for detected credentials, if
9427
+ # such user input is found.
9428
+ #
9386
9429
  #
9387
9430
  #
9388
9431
  # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html
@@ -17694,7 +17737,7 @@ module Aws::SageMaker
17694
17737
  #
17695
17738
  # @!attribute [rw] execution_role_identity_config
17696
17739
  # The configuration for attaching a SageMaker user profile name to the
17697
- # execution role as a [ `sts:SourceIdentity` key][1].
17740
+ # execution role as a [sts:SourceIdentity key][1].
17698
17741
  #
17699
17742
  #
17700
17743
  #
@@ -17736,9 +17779,9 @@ module Aws::SageMaker
17736
17779
  #
17737
17780
  # @!attribute [rw] execution_role_identity_config
17738
17781
  # The configuration for attaching a SageMaker user profile name to the
17739
- # execution role as a [ `sts:SourceIdentity` key][1]. This
17740
- # configuration can only be modified if there are no apps in the
17741
- # `InService` or `Pending` state.
17782
+ # execution role as a [sts:SourceIdentity key][1]. This configuration
17783
+ # can only be modified if there are no apps in the `InService` or
17784
+ # `Pending` state.
17742
17785
  #
17743
17786
  #
17744
17787
  #
@@ -21141,7 +21184,8 @@ module Aws::SageMaker
21141
21184
  # Defines the maximum number of data objects that can be labeled by
21142
21185
  # human workers at the same time. Also referred to as batch size. Each
21143
21186
  # object may have more than one worker at one time. The default value
21144
- # is 1000 objects.
21187
+ # is 1000 objects. To increase the maximum value to 5000 objects,
21188
+ # contact Amazon Web Services Support.
21145
21189
  # @return [Integer]
21146
21190
  #
21147
21191
  # @!attribute [rw] annotation_consolidation_config
@@ -21810,7 +21854,13 @@ module Aws::SageMaker
21810
21854
  # data as a hash:
21811
21855
  #
21812
21856
  # {
21813
- # strategy: "Bayesian", # required, accepts Bayesian, Random
21857
+ # strategy: "Bayesian", # required, accepts Bayesian, Random, Hyperband
21858
+ # strategy_config: {
21859
+ # hyperband_strategy_config: {
21860
+ # min_resource: 1,
21861
+ # max_resource: 1,
21862
+ # },
21863
+ # },
21814
21864
  # hyper_parameter_tuning_job_objective: {
21815
21865
  # type: "Maximize", # required, accepts Maximize, Minimize
21816
21866
  # metric_name: "MetricName", # required
@@ -21851,16 +21901,21 @@ module Aws::SageMaker
21851
21901
  #
21852
21902
  # @!attribute [rw] strategy
21853
21903
  # Specifies how hyperparameter tuning chooses the combinations of
21854
- # hyperparameter values to use for the training job it launches. To
21855
- # use the Bayesian search strategy, set this to `Bayesian`. To
21856
- # randomly search, set it to `Random`. For information about search
21857
- # strategies, see [How Hyperparameter Tuning Works][1].
21904
+ # hyperparameter values to use for the training job it launches. For
21905
+ # information about search strategies, see [How Hyperparameter Tuning
21906
+ # Works][1].
21858
21907
  #
21859
21908
  #
21860
21909
  #
21861
21910
  # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/automatic-model-tuning-how-it-works.html
21862
21911
  # @return [String]
21863
21912
  #
21913
+ # @!attribute [rw] strategy_config
21914
+ # The configuration for the `Hyperband` optimization strategy. This
21915
+ # parameter should be provided only if `Hyperband` is selected as the
21916
+ # strategy for `HyperParameterTuningJobConfig`.
21917
+ # @return [Types::HyperParameterTuningJobStrategyConfig]
21918
+ #
21864
21919
  # @!attribute [rw] hyper_parameter_tuning_job_objective
21865
21920
  # The HyperParameterTuningJobObjective object that specifies the
21866
21921
  # objective metric for this tuning job.
@@ -21878,8 +21933,11 @@ module Aws::SageMaker
21878
21933
  #
21879
21934
  # @!attribute [rw] training_job_early_stopping_type
21880
21935
  # Specifies whether to use early stopping for training jobs launched
21881
- # by the hyperparameter tuning job. This can be one of the following
21882
- # values (the default value is `OFF`):
21936
+ # by the hyperparameter tuning job. Because the `Hyperband` strategy
21937
+ # has its own advanced internal early stopping mechanism,
21938
+ # `TrainingJobEarlyStoppingType` must be `OFF` to use `Hyperband`.
21939
+ # This parameter can take on one of the following values (the default
21940
+ # value is `OFF`):
21883
21941
  #
21884
21942
  # OFF
21885
21943
  #
@@ -21906,6 +21964,7 @@ module Aws::SageMaker
21906
21964
  #
21907
21965
  class HyperParameterTuningJobConfig < Struct.new(
21908
21966
  :strategy,
21967
+ :strategy_config,
21909
21968
  :hyper_parameter_tuning_job_objective,
21910
21969
  :resource_limits,
21911
21970
  :parameter_ranges,
@@ -22063,6 +22122,42 @@ module Aws::SageMaker
22063
22122
  include Aws::Structure
22064
22123
  end
22065
22124
 
22125
+ # The configuration for a training job launched by a hyperparameter
22126
+ # tuning job. Choose `Bayesian` for Bayesian optimization, and `Random`
22127
+ # for random search optimization. For more advanced use cases, use
22128
+ # `Hyperband`, which evaluates objective metrics for training jobs after
22129
+ # every epoch. For more information about strategies, see [How
22130
+ # Hyperparameter Tuning Works][1].
22131
+ #
22132
+ #
22133
+ #
22134
+ # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/automatic-model-tuning-how-it-works.html
22135
+ #
22136
+ # @note When making an API call, you may pass HyperParameterTuningJobStrategyConfig
22137
+ # data as a hash:
22138
+ #
22139
+ # {
22140
+ # hyperband_strategy_config: {
22141
+ # min_resource: 1,
22142
+ # max_resource: 1,
22143
+ # },
22144
+ # }
22145
+ #
22146
+ # @!attribute [rw] hyperband_strategy_config
22147
+ # The configuration for the object that specifies the `Hyperband`
22148
+ # strategy. This parameter is only supported for the `Hyperband`
22149
+ # selection for `Strategy` within the `HyperParameterTuningJobConfig`
22150
+ # API.
22151
+ # @return [Types::HyperbandStrategyConfig]
22152
+ #
22153
+ # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/HyperParameterTuningJobStrategyConfig AWS API Documentation
22154
+ #
22155
+ class HyperParameterTuningJobStrategyConfig < Struct.new(
22156
+ :hyperband_strategy_config)
22157
+ SENSITIVE = []
22158
+ include Aws::Structure
22159
+ end
22160
+
22066
22161
  # Provides summary information about a hyperparameter tuning job.
22067
22162
  #
22068
22163
  # @!attribute [rw] hyper_parameter_tuning_job_name
@@ -22079,8 +22174,7 @@ module Aws::SageMaker
22079
22174
  #
22080
22175
  # @!attribute [rw] strategy
22081
22176
  # Specifies the search strategy hyperparameter tuning uses to choose
22082
- # which hyperparameters to use for each iteration. Currently, the only
22083
- # valid value is Bayesian.
22177
+ # which hyperparameters to evaluate at each iteration.
22084
22178
  # @return [String]
22085
22179
  #
22086
22180
  # @!attribute [rw] creation_time
@@ -22348,6 +22442,74 @@ module Aws::SageMaker
22348
22442
  include Aws::Structure
22349
22443
  end
22350
22444
 
22445
+ # The configuration for `Hyperband`, a multi-fidelity based
22446
+ # hyperparameter tuning strategy. `Hyperband` uses the final and
22447
+ # intermediate results of a training job to dynamically allocate
22448
+ # resources to utilized hyperparameter configurations while
22449
+ # automatically stopping under-performing configurations. This parameter
22450
+ # should be provided only if `Hyperband` is selected as the
22451
+ # `StrategyConfig` under the `HyperParameterTuningJobConfig` API.
22452
+ #
22453
+ # @note When making an API call, you may pass HyperbandStrategyConfig
22454
+ # data as a hash:
22455
+ #
22456
+ # {
22457
+ # min_resource: 1,
22458
+ # max_resource: 1,
22459
+ # }
22460
+ #
22461
+ # @!attribute [rw] min_resource
22462
+ # The minimum number of resources (such as epochs) that can be used by
22463
+ # a training job launched by a hyperparameter tuning job. If the value
22464
+ # for `MinResource` has not been reached, the training job will not be
22465
+ # stopped by `Hyperband`.
22466
+ # @return [Integer]
22467
+ #
22468
+ # @!attribute [rw] max_resource
22469
+ # The maximum number of resources (such as epochs) that can be used by
22470
+ # a training job launched by a hyperparameter tuning job. Once a job
22471
+ # reaches the `MaxResource` value, it is stopped. If a value for
22472
+ # `MaxResource` is not provided, and `Hyperband` is selected as the
22473
+ # hyperparameter tuning strategy, `HyperbandTrainingJ` attempts to
22474
+ # infer `MaxResource` from the following keys (if present) in
22475
+ # [StaticsHyperParameters][1]\:
22476
+ #
22477
+ # * `epochs`
22478
+ #
22479
+ # * `numepochs`
22480
+ #
22481
+ # * `n-epochs`
22482
+ #
22483
+ # * `n_epochs`
22484
+ #
22485
+ # * `num_epochs`
22486
+ #
22487
+ # If `HyperbandStrategyConfig` is unable to infer a value for
22488
+ # `MaxResource`, it generates a validation error. The maximum value is
22489
+ # 20,000 epochs. All metrics that correspond to an objective metric
22490
+ # are used to derive [early stopping decisions][2]. For
22491
+ # [distributive][3] training jobs, ensure that duplicate metrics are
22492
+ # not printed in the logs across the individual nodes in a training
22493
+ # job. If multiple nodes are publishing duplicate or incorrect
22494
+ # metrics, training jobs may make an incorrect stopping decision and
22495
+ # stop the job prematurely.
22496
+ #
22497
+ #
22498
+ #
22499
+ # [1]: https://docs.aws.amazon.com/sagemaker/latest/APIReference/API_HyperParameterTrainingJobDefinition.html#sagemaker-Type-HyperParameterTrainingJobDefinition-StaticHyperParameters
22500
+ # [2]: https://docs.aws.amazon.com/sagemaker/latest/dg/automatic-model-tuning-early-stopping.html
22501
+ # [3]: https://docs.aws.amazon.com/sagemaker/latest/dg/distributed-training.html
22502
+ # @return [Integer]
22503
+ #
22504
+ # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/HyperbandStrategyConfig AWS API Documentation
22505
+ #
22506
+ class HyperbandStrategyConfig < Struct.new(
22507
+ :min_resource,
22508
+ :max_resource)
22509
+ SENSITIVE = []
22510
+ include Aws::Structure
22511
+ end
22512
+
22351
22513
  # A SageMaker image. A SageMaker image represents a set of container
22352
22514
  # images that are derived from a common base container image. Each of
22353
22515
  # these container images is represented by a SageMaker `ImageVersion`.
@@ -34097,16 +34259,16 @@ module Aws::SageMaker
34097
34259
  # @return [Integer]
34098
34260
  #
34099
34261
  # @!attribute [rw] model_data_download_timeout_in_seconds
34100
- # The timeout value, in seconds, to download and extract customer
34101
- # model artifact from Amazon S3 to individual inference instance
34102
- # associated with this production variant.
34262
+ # The timeout value, in seconds, to download and extract the model
34263
+ # that you want to host from Amazon S3 to the individual inference
34264
+ # instance associated with this production variant.
34103
34265
  # @return [Integer]
34104
34266
  #
34105
34267
  # @!attribute [rw] container_startup_health_check_timeout_in_seconds
34106
- # The timeout value, in seconds, for the customer inference container
34107
- # to pass health check by SageMaker Hosting. For more information on
34108
- # health check, see [How Your Container Should Respond to Health Check
34109
- # (Ping) Requests][1].
34268
+ # The timeout value, in seconds, for your inference container to pass
34269
+ # health check by SageMaker Hosting. For more information about health
34270
+ # check, see [How Your Container Should Respond to Health Check (Ping)
34271
+ # Requests][1].
34110
34272
  #
34111
34273
  #
34112
34274
  #
@@ -49,6 +49,6 @@ require_relative 'aws-sdk-sagemaker/customizations'
49
49
  # @!group service
50
50
  module Aws::SageMaker
51
51
 
52
- GEM_VERSION = '1.138.0'
52
+ GEM_VERSION = '1.140.0'
53
53
 
54
54
  end
metadata CHANGED
@@ -1,14 +1,14 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: aws-sdk-sagemaker
3
3
  version: !ruby/object:Gem::Version
4
- version: 1.138.0
4
+ version: 1.140.0
5
5
  platform: ruby
6
6
  authors:
7
7
  - Amazon Web Services
8
8
  autorequire:
9
9
  bindir: bin
10
10
  cert_chain: []
11
- date: 2022-09-07 00:00:00.000000000 Z
11
+ date: 2022-09-15 00:00:00.000000000 Z
12
12
  dependencies:
13
13
  - !ruby/object:Gem::Dependency
14
14
  name: aws-sdk-core