diff --git a/clients/client-sagemaker/src/commands/CreateInferenceRecommendationsJobCommand.ts b/clients/client-sagemaker/src/commands/CreateInferenceRecommendationsJobCommand.ts
index dc8b4d679981..418794f8ec73 100644
--- a/clients/client-sagemaker/src/commands/CreateInferenceRecommendationsJobCommand.ts
+++ b/clients/client-sagemaker/src/commands/CreateInferenceRecommendationsJobCommand.ts
@@ -75,7 +75,7 @@ export interface CreateInferenceRecommendationsJobCommandOutput
* },
* EndpointConfigurations: [ // EndpointInputConfigurations
* { // EndpointInputConfiguration
- * InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge", // required
+ * InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge",
* InferenceSpecificationName: "STRING_VALUE",
* EnvironmentParameterRanges: { // EnvironmentParameterRanges
* CategoricalParameterRanges: [ // CategoricalParameters
@@ -87,6 +87,11 @@ export interface CreateInferenceRecommendationsJobCommandOutput
* },
* ],
* },
+ * ServerlessConfig: { // ProductionVariantServerlessConfig
+ * MemorySizeInMB: Number("int"), // required
+ * MaxConcurrency: Number("int"), // required
+ * ProvisionedConcurrency: Number("int"),
+ * },
* },
* ],
* VolumeKmsKeyId: "STRING_VALUE",
@@ -106,6 +111,7 @@ export interface CreateInferenceRecommendationsJobCommandOutput
* "STRING_VALUE",
* ],
* DataInputConfig: "STRING_VALUE",
+ * SupportedEndpointType: "RealTime" || "Serverless",
* },
* Endpoints: [ // Endpoints
* { // EndpointInfo
diff --git a/clients/client-sagemaker/src/commands/DeleteHumanTaskUiCommand.ts b/clients/client-sagemaker/src/commands/DeleteHumanTaskUiCommand.ts
index 0afafbf7c27a..b5588a2e442c 100644
--- a/clients/client-sagemaker/src/commands/DeleteHumanTaskUiCommand.ts
+++ b/clients/client-sagemaker/src/commands/DeleteHumanTaskUiCommand.ts
@@ -13,7 +13,8 @@ import {
SerdeContext as __SerdeContext,
} from "@smithy/types";
-import { DeleteHumanTaskUiRequest, DeleteHumanTaskUiResponse } from "../models/models_1";
+import { DeleteHumanTaskUiRequest } from "../models/models_1";
+import { DeleteHumanTaskUiResponse } from "../models/models_2";
import { de_DeleteHumanTaskUiCommand, se_DeleteHumanTaskUiCommand } from "../protocols/Aws_json1_1";
import { SageMakerClientResolvedConfig, ServiceInputTypes, ServiceOutputTypes } from "../SageMakerClient";
diff --git a/clients/client-sagemaker/src/commands/DescribeInferenceRecommendationsJobCommand.ts b/clients/client-sagemaker/src/commands/DescribeInferenceRecommendationsJobCommand.ts
index 0de41487f496..92856acd21d7 100644
--- a/clients/client-sagemaker/src/commands/DescribeInferenceRecommendationsJobCommand.ts
+++ b/clients/client-sagemaker/src/commands/DescribeInferenceRecommendationsJobCommand.ts
@@ -87,7 +87,7 @@ export interface DescribeInferenceRecommendationsJobCommandOutput
* // },
* // EndpointConfigurations: [ // EndpointInputConfigurations
* // { // EndpointInputConfiguration
- * // InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge", // required
+ * // InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge",
* // InferenceSpecificationName: "STRING_VALUE",
* // EnvironmentParameterRanges: { // EnvironmentParameterRanges
* // CategoricalParameterRanges: [ // CategoricalParameters
@@ -99,6 +99,11 @@ export interface DescribeInferenceRecommendationsJobCommandOutput
* // },
* // ],
* // },
+ * // ServerlessConfig: { // ProductionVariantServerlessConfig
+ * // MemorySizeInMB: Number("int"), // required
+ * // MaxConcurrency: Number("int"), // required
+ * // ProvisionedConcurrency: Number("int"),
+ * // },
* // },
* // ],
* // VolumeKmsKeyId: "STRING_VALUE",
@@ -118,6 +123,7 @@ export interface DescribeInferenceRecommendationsJobCommandOutput
* // "STRING_VALUE",
* // ],
* // DataInputConfig: "STRING_VALUE",
+ * // SupportedEndpointType: "RealTime" || "Serverless",
* // },
* // Endpoints: [ // Endpoints
* // { // EndpointInfo
@@ -152,12 +158,18 @@ export interface DescribeInferenceRecommendationsJobCommandOutput
* // ModelLatency: Number("int"), // required
* // CpuUtilization: Number("float"),
* // MemoryUtilization: Number("float"),
+ * // ModelSetupTime: Number("int"),
* // },
* // EndpointConfiguration: { // EndpointOutputConfiguration
* // EndpointName: "STRING_VALUE", // required
* // VariantName: "STRING_VALUE", // required
- * // InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge", // required
- * // InitialInstanceCount: Number("int"), // required
+ * // InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge",
+ * // InitialInstanceCount: Number("int"),
+ * // ServerlessConfig: {
+ * // MemorySizeInMB: Number("int"), // required
+ * // MaxConcurrency: Number("int"), // required
+ * // ProvisionedConcurrency: Number("int"),
+ * // },
* // },
* // ModelConfiguration: { // ModelConfiguration
* // InferenceSpecificationName: "STRING_VALUE",
diff --git a/clients/client-sagemaker/src/commands/ListInferenceRecommendationsJobStepsCommand.ts b/clients/client-sagemaker/src/commands/ListInferenceRecommendationsJobStepsCommand.ts
index 2416b4fab51f..579d6efb6777 100644
--- a/clients/client-sagemaker/src/commands/ListInferenceRecommendationsJobStepsCommand.ts
+++ b/clients/client-sagemaker/src/commands/ListInferenceRecommendationsJobStepsCommand.ts
@@ -75,12 +75,18 @@ export interface ListInferenceRecommendationsJobStepsCommandOutput
* // ModelLatency: Number("int"), // required
* // CpuUtilization: Number("float"),
* // MemoryUtilization: Number("float"),
+ * // ModelSetupTime: Number("int"),
* // },
* // EndpointConfiguration: { // EndpointOutputConfiguration
* // EndpointName: "STRING_VALUE", // required
* // VariantName: "STRING_VALUE", // required
- * // InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge", // required
- * // InitialInstanceCount: Number("int"), // required
+ * // InstanceType: "ml.t2.medium" || "ml.t2.large" || "ml.t2.xlarge" || "ml.t2.2xlarge" || "ml.m4.xlarge" || "ml.m4.2xlarge" || "ml.m4.4xlarge" || "ml.m4.10xlarge" || "ml.m4.16xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.12xlarge" || "ml.m5.24xlarge" || "ml.m5d.large" || "ml.m5d.xlarge" || "ml.m5d.2xlarge" || "ml.m5d.4xlarge" || "ml.m5d.12xlarge" || "ml.m5d.24xlarge" || "ml.c4.large" || "ml.c4.xlarge" || "ml.c4.2xlarge" || "ml.c4.4xlarge" || "ml.c4.8xlarge" || "ml.p2.xlarge" || "ml.p2.8xlarge" || "ml.p2.16xlarge" || "ml.p3.2xlarge" || "ml.p3.8xlarge" || "ml.p3.16xlarge" || "ml.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.18xlarge" || "ml.c5d.large" || "ml.c5d.xlarge" || "ml.c5d.2xlarge" || "ml.c5d.4xlarge" || "ml.c5d.9xlarge" || "ml.c5d.18xlarge" || "ml.g4dn.xlarge" || "ml.g4dn.2xlarge" || "ml.g4dn.4xlarge" || "ml.g4dn.8xlarge" || "ml.g4dn.12xlarge" || "ml.g4dn.16xlarge" || "ml.r5.large" || "ml.r5.xlarge" || "ml.r5.2xlarge" || "ml.r5.4xlarge" || "ml.r5.12xlarge" || "ml.r5.24xlarge" || "ml.r5d.large" || "ml.r5d.xlarge" || "ml.r5d.2xlarge" || "ml.r5d.4xlarge" || "ml.r5d.12xlarge" || "ml.r5d.24xlarge" || "ml.inf1.xlarge" || "ml.inf1.2xlarge" || "ml.inf1.6xlarge" || "ml.inf1.24xlarge" || "ml.c6i.large" || "ml.c6i.xlarge" || "ml.c6i.2xlarge" || "ml.c6i.4xlarge" || "ml.c6i.8xlarge" || "ml.c6i.12xlarge" || "ml.c6i.16xlarge" || "ml.c6i.24xlarge" || "ml.c6i.32xlarge" || "ml.g5.xlarge" || "ml.g5.2xlarge" || "ml.g5.4xlarge" || "ml.g5.8xlarge" || "ml.g5.12xlarge" || "ml.g5.16xlarge" || "ml.g5.24xlarge" || "ml.g5.48xlarge" || "ml.p4d.24xlarge" || "ml.c7g.large" || "ml.c7g.xlarge" || "ml.c7g.2xlarge" || "ml.c7g.4xlarge" || "ml.c7g.8xlarge" || "ml.c7g.12xlarge" || "ml.c7g.16xlarge" || "ml.m6g.large" || "ml.m6g.xlarge" || "ml.m6g.2xlarge" || "ml.m6g.4xlarge" || "ml.m6g.8xlarge" || "ml.m6g.12xlarge" || "ml.m6g.16xlarge" || "ml.m6gd.large" || "ml.m6gd.xlarge" || "ml.m6gd.2xlarge" || "ml.m6gd.4xlarge" || "ml.m6gd.8xlarge" || "ml.m6gd.12xlarge" || "ml.m6gd.16xlarge" || "ml.c6g.large" || "ml.c6g.xlarge" || "ml.c6g.2xlarge" || "ml.c6g.4xlarge" || "ml.c6g.8xlarge" || "ml.c6g.12xlarge" || "ml.c6g.16xlarge" || "ml.c6gd.large" || "ml.c6gd.xlarge" || "ml.c6gd.2xlarge" || "ml.c6gd.4xlarge" || "ml.c6gd.8xlarge" || "ml.c6gd.12xlarge" || "ml.c6gd.16xlarge" || "ml.c6gn.large" || "ml.c6gn.xlarge" || "ml.c6gn.2xlarge" || "ml.c6gn.4xlarge" || "ml.c6gn.8xlarge" || "ml.c6gn.12xlarge" || "ml.c6gn.16xlarge" || "ml.r6g.large" || "ml.r6g.xlarge" || "ml.r6g.2xlarge" || "ml.r6g.4xlarge" || "ml.r6g.8xlarge" || "ml.r6g.12xlarge" || "ml.r6g.16xlarge" || "ml.r6gd.large" || "ml.r6gd.xlarge" || "ml.r6gd.2xlarge" || "ml.r6gd.4xlarge" || "ml.r6gd.8xlarge" || "ml.r6gd.12xlarge" || "ml.r6gd.16xlarge" || "ml.p4de.24xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge",
+ * // InitialInstanceCount: Number("int"),
+ * // ServerlessConfig: { // ProductionVariantServerlessConfig
+ * // MemorySizeInMB: Number("int"), // required
+ * // MaxConcurrency: Number("int"), // required
+ * // ProvisionedConcurrency: Number("int"),
+ * // },
* // },
* // ModelConfiguration: { // ModelConfiguration
* // InferenceSpecificationName: "STRING_VALUE",
diff --git a/clients/client-sagemaker/src/models/models_0.ts b/clients/client-sagemaker/src/models/models_0.ts
index 5dc00b5129a0..2ce6f348f70b 100644
--- a/clients/client-sagemaker/src/models/models_0.ts
+++ b/clients/client-sagemaker/src/models/models_0.ts
@@ -5902,6 +5902,11 @@ export namespace AutoMLProblemTypeConfig {
/**
* Settings used to configure an AutoML job V2 for a time-series forecasting problem
* type.
+ *
+ * The TimeSeriesForecastingJobConfig problem type is only available in
+ * private beta. Contact Amazon Web Services Support or your account manager to learn more
+ * about access privileges.
+ *
*/
export interface TimeSeriesForecastingJobConfigMember {
ImageClassificationJobConfig?: never;
diff --git a/clients/client-sagemaker/src/models/models_1.ts b/clients/client-sagemaker/src/models/models_1.ts
index 14bbd21dba60..14afc7bce049 100644
--- a/clients/client-sagemaker/src/models/models_1.ts
+++ b/clients/client-sagemaker/src/models/models_1.ts
@@ -605,9 +605,10 @@ export interface CreateEdgePackagingJobRequest {
*/
export interface RollingUpdatePolicy {
/**
- * Batch size for each rolling step to provision capacity and turn on traffic on the new
- * endpoint fleet, and terminate capacity on the old endpoint fleet. Value must be between
- * 5% to 50% of the variant's total instance count.
+ * Specifies the type and size of the endpoint capacity to activate for a blue/green deployment, a rolling deployment, or a rollback strategy.
+ * You can specify your batches as either instance count or the overall percentage or your fleet.
+ * For a rollback strategy, if you don't specify the fields in this object, or if you set the Value to 100%, then SageMaker
+ * uses a blue/green rollback strategy and rolls all traffic back to the blue fleet.
*/
MaximumBatchSize: CapacitySize | undefined;
@@ -622,10 +623,10 @@ export interface RollingUpdatePolicy {
MaximumExecutionTimeoutInSeconds?: number;
/**
- * Batch size for rollback to the old endpoint fleet. Each rolling step to provision
- * capacity and turn on traffic on the old endpoint fleet, and terminate capacity on the new
- * endpoint fleet. If this field is absent, the default value will be set to 100% of total
- * capacity which means to bring up the whole capacity of the old fleet at once during rollback.
+ * Specifies the type and size of the endpoint capacity to activate for a blue/green deployment, a rolling deployment, or a rollback strategy.
+ * You can specify your batches as either instance count or the overall percentage or your fleet.
+ * For a rollback strategy, if you don't specify the fields in this object, or if you set the Value to 100%, then SageMaker
+ * uses a blue/green rollback strategy and rolls all traffic back to the blue fleet.
*/
RollbackMaximumBatchSize?: CapacitySize;
}
@@ -867,6 +868,11 @@ export interface ProductionVariantServerlessConfig {
/**
* The amount of provisioned concurrency to allocate for the serverless endpoint.
* Should be less than or equal to MaxConcurrency.
+ *
+ * This field is not supported for serverless endpoint recommendations for Inference Recommender jobs.
+ * For more information about creating an Inference Recommender job, see
+ * CreateInferenceRecommendationsJobs.
+ *
*/
ProvisionedConcurrency?: number;
}
@@ -3986,6 +3992,21 @@ export interface RecommendationJobPayloadConfig {
SupportedContentTypes?: string[];
}
+/**
+ * @public
+ * @enum
+ */
+export const RecommendationJobSupportedEndpointType = {
+ REALTIME: "RealTime",
+ SERVERLESS: "Serverless",
+} as const;
+
+/**
+ * @public
+ */
+export type RecommendationJobSupportedEndpointType =
+ (typeof RecommendationJobSupportedEndpointType)[keyof typeof RecommendationJobSupportedEndpointType];
+
/**
* @public
* Specifies mandatory fields for running an Inference Recommender job directly in the
@@ -4047,6 +4068,13 @@ export interface RecommendationJobContainerConfig {
* DataInputConfig.
*/
DataInputConfig?: string;
+
+ /**
+ * The endpoint type to receive recommendations for. By default this is null, and the results of
+ * the inference recommendation job return a combined list of both real-time and serverless benchmarks.
+ * By specifying a value for this field, you can receive a longer list of benchmarks for the desired endpoint type.
+ */
+ SupportedEndpointType?: RecommendationJobSupportedEndpointType | string;
}
/**
@@ -4068,7 +4096,7 @@ export interface EndpointInputConfiguration {
/**
* The instance types to use for the load test.
*/
- InstanceType: ProductionVariantInstanceType | string | undefined;
+ InstanceType?: ProductionVariantInstanceType | string;
/**
* The inference specification name in the model package version.
@@ -4079,6 +4107,11 @@ export interface EndpointInputConfiguration {
* The parameter you want to benchmark against.
*/
EnvironmentParameterRanges?: EnvironmentParameterRanges;
+
+ /**
+ * Specifies the serverless configuration for an endpoint variant.
+ */
+ ServerlessConfig?: ProductionVariantServerlessConfig;
}
/**
@@ -10786,11 +10819,6 @@ export interface DeleteHumanTaskUiRequest {
HumanTaskUiName: string | undefined;
}
-/**
- * @public
- */
-export interface DeleteHumanTaskUiResponse {}
-
/**
* @internal
*/
diff --git a/clients/client-sagemaker/src/models/models_2.ts b/clients/client-sagemaker/src/models/models_2.ts
index 50d0c7d14b8f..8f63258f2280 100644
--- a/clients/client-sagemaker/src/models/models_2.ts
+++ b/clients/client-sagemaker/src/models/models_2.ts
@@ -177,6 +177,11 @@ import {
VendorGuidance,
} from "./models_1";
+/**
+ * @public
+ */
+export interface DeleteHumanTaskUiResponse {}
+
/**
* @public
*/
@@ -4430,12 +4435,17 @@ export interface EndpointOutputConfiguration {
/**
* The instance type recommended by Amazon SageMaker Inference Recommender.
*/
- InstanceType: ProductionVariantInstanceType | string | undefined;
+ InstanceType?: ProductionVariantInstanceType | string;
/**
* The number of instances recommended to launch initially.
*/
- InitialInstanceCount: number | undefined;
+ InitialInstanceCount?: number;
+
+ /**
+ * Specifies the serverless configuration for an endpoint variant.
+ */
+ ServerlessConfig?: ProductionVariantServerlessConfig;
}
/**
@@ -4476,6 +4486,15 @@ export interface RecommendationMetrics {
* NaN indicates that the value is not available.
*/
MemoryUtilization?: number;
+
+ /**
+ * The time it takes to launch new compute resources for a serverless endpoint.
+ * The time can vary depending on the model size, how long it takes to download the
+ * model, and the start-up time of the container.
+ *
+ * NaN indicates that the value is not available.
+ */
+ ModelSetupTime?: number;
}
/**
@@ -10093,18 +10112,6 @@ export const ResourceType = {
*/
export type ResourceType = (typeof ResourceType)[keyof typeof ResourceType];
-/**
- * @public
- * Part of the SuggestionQuery type. Specifies a hint for retrieving property
- * names that begin with the specified text.
- */
-export interface PropertyNameQuery {
- /**
- * Text that begins a property's name.
- */
- PropertyNameHint: string | undefined;
-}
-
/**
* @internal
*/
diff --git a/clients/client-sagemaker/src/models/models_3.ts b/clients/client-sagemaker/src/models/models_3.ts
index d60c4f5ce863..f8a4ca341ac0 100644
--- a/clients/client-sagemaker/src/models/models_3.ts
+++ b/clients/client-sagemaker/src/models/models_3.ts
@@ -89,8 +89,6 @@ import {
SourceAlgorithmSpecification,
StudioLifecycleConfigAppType,
TensorBoardOutputConfig,
- TrialComponentArtifact,
- TrialComponentParameterValue,
TrialComponentStatus,
TtlDuration,
UiTemplate,
@@ -151,7 +149,6 @@ import {
PipelineStatus,
ProcessingJobStatus,
ProjectStatus,
- PropertyNameQuery,
RecommendationJobStatus,
RecommendationMetrics,
ResourceType,
@@ -165,7 +162,6 @@ import {
TrainingJobStatus,
TrainingJobStatusCounters,
TransformJobStatus,
- TrialComponentMetricSummary,
TrialComponentSource,
TrialSource,
UserProfileStatus,
@@ -175,6 +171,18 @@ import {
Workteam,
} from "./models_2";
+/**
+ * @public
+ * Part of the SuggestionQuery type. Specifies a hint for retrieving property
+ * names that begin with the specified text.
+ */
+export interface PropertyNameQuery {
+ /**
+ * Text that begins a property's name.
+ */
+ PropertyNameHint: string | undefined;
+}
+
/**
* @public
* Specified in the GetSearchSuggestions request.
@@ -10039,121 +10047,6 @@ export interface TrialComponentSourceDetail {
TransformJob?: TransformJob;
}
-/**
- * @public
- *
The properties of a trial component as returned by the Search
- * API.
- */
-export interface TrialComponent {
- /**
- * The name of the trial component.
- */
- TrialComponentName?: string;
-
- /**
- * The name of the component as displayed. If DisplayName isn't specified,
- * TrialComponentName is displayed.
- */
- DisplayName?: string;
-
- /**
- * The Amazon Resource Name (ARN) of the trial component.
- */
- TrialComponentArn?: string;
-
- /**
- * The Amazon Resource Name (ARN) and job type of the source of the component.
- */
- Source?: TrialComponentSource;
-
- /**
- * The status of the trial component.
- */
- Status?: TrialComponentStatus;
-
- /**
- * When the component started.
- */
- StartTime?: Date;
-
- /**
- * When the component ended.
- */
- EndTime?: Date;
-
- /**
- * When the component was created.
- */
- CreationTime?: Date;
-
- /**
- * Who created the trial component.
- */
- CreatedBy?: UserContext;
-
- /**
- * When the component was last modified.
- */
- LastModifiedTime?: Date;
-
- /**
- * Information about the user who created or modified an experiment, trial, trial
- * component, lineage group, project, or model card.
- */
- LastModifiedBy?: UserContext;
-
- /**
- * The hyperparameters of the component.
- */
- Parameters?: Record;
-
- /**
- * The input artifacts of the component.
- */
- InputArtifacts?: Record;
-
- /**
- * The output artifacts of the component.
- */
- OutputArtifacts?: Record;
-
- /**
- * The metrics for the component.
- */
- Metrics?: TrialComponentMetricSummary[];
-
- /**
- * Metadata properties of the tracking entity, trial, or trial component.
- */
- MetadataProperties?: MetadataProperties;
-
- /**
- * Details of the source of the component.
- */
- SourceDetail?: TrialComponentSourceDetail;
-
- /**
- * The Amazon Resource Name (ARN) of the lineage group resource.
- */
- LineageGroupArn?: string;
-
- /**
- * The list of tags that are associated with the component. You can use Search API to search on the tags.
- */
- Tags?: Tag[];
-
- /**
- * An array of the parents of the component. A parent is a trial the component is associated
- * with and the experiment the trial is part of. A component might not have any parents.
- */
- Parents?: Parent[];
-
- /**
- * The name of the experiment run.
- */
- RunName?: string;
-}
-
/**
* @internal
*/
diff --git a/clients/client-sagemaker/src/models/models_4.ts b/clients/client-sagemaker/src/models/models_4.ts
index 4e78f0dd7af7..f56294cf5e25 100644
--- a/clients/client-sagemaker/src/models/models_4.ts
+++ b/clients/client-sagemaker/src/models/models_4.ts
@@ -9,9 +9,11 @@ import {
DefaultSpaceSettings,
EdgeOutputConfig,
KernelGatewayImageConfig,
+ MetadataProperties,
ModelApprovalStatus,
OutputParameter,
Tag,
+ UserContext,
} from "./models_0";
import {
_InstanceType,
@@ -58,6 +60,8 @@ import {
Filter,
ResourceType,
SelectiveExecutionConfig,
+ TrialComponentMetricSummary,
+ TrialComponentSource,
Workforce,
Workteam,
} from "./models_2";
@@ -74,6 +78,7 @@ import {
NestedFilters,
OnlineStoreConfigUpdate,
Parameter,
+ Parent,
Pipeline,
PipelineExecution,
ProfilerConfigForUpdate,
@@ -82,9 +87,124 @@ import {
SearchSortOrder,
TrainingJob,
Trial,
- TrialComponent,
+ TrialComponentSourceDetail,
} from "./models_3";
+/**
+ * @public
+ * The properties of a trial component as returned by the Search
+ * API.
+ */
+export interface TrialComponent {
+ /**
+ * The name of the trial component.
+ */
+ TrialComponentName?: string;
+
+ /**
+ * The name of the component as displayed. If DisplayName isn't specified,
+ * TrialComponentName is displayed.
+ */
+ DisplayName?: string;
+
+ /**
+ * The Amazon Resource Name (ARN) of the trial component.
+ */
+ TrialComponentArn?: string;
+
+ /**
+ * The Amazon Resource Name (ARN) and job type of the source of the component.
+ */
+ Source?: TrialComponentSource;
+
+ /**
+ * The status of the trial component.
+ */
+ Status?: TrialComponentStatus;
+
+ /**
+ * When the component started.
+ */
+ StartTime?: Date;
+
+ /**
+ * When the component ended.
+ */
+ EndTime?: Date;
+
+ /**
+ * When the component was created.
+ */
+ CreationTime?: Date;
+
+ /**
+ * Who created the trial component.
+ */
+ CreatedBy?: UserContext;
+
+ /**
+ * When the component was last modified.
+ */
+ LastModifiedTime?: Date;
+
+ /**
+ * Information about the user who created or modified an experiment, trial, trial
+ * component, lineage group, project, or model card.
+ */
+ LastModifiedBy?: UserContext;
+
+ /**
+ * The hyperparameters of the component.
+ */
+ Parameters?: Record;
+
+ /**
+ * The input artifacts of the component.
+ */
+ InputArtifacts?: Record;
+
+ /**
+ * The output artifacts of the component.
+ */
+ OutputArtifacts?: Record;
+
+ /**
+ * The metrics for the component.
+ */
+ Metrics?: TrialComponentMetricSummary[];
+
+ /**
+ * Metadata properties of the tracking entity, trial, or trial component.
+ */
+ MetadataProperties?: MetadataProperties;
+
+ /**
+ * Details of the source of the component.
+ */
+ SourceDetail?: TrialComponentSourceDetail;
+
+ /**
+ * The Amazon Resource Name (ARN) of the lineage group resource.
+ */
+ LineageGroupArn?: string;
+
+ /**
+ * The list of tags that are associated with the component. You can use Search API to search on the tags.
+ */
+ Tags?: Tag[];
+
+ /**
+ * An array of the parents of the component. A parent is a trial the component is associated
+ * with and the experiment the trial is part of. A component might not have any parents.
+ */
+ Parents?: Parent[];
+
+ /**
+ * The name of the experiment run.
+ */
+ RunName?: string;
+}
+
/**
* @public
* A single resource returned as part of the Search API response.
diff --git a/clients/client-sagemaker/src/protocols/Aws_json1_1.ts b/clients/client-sagemaker/src/protocols/Aws_json1_1.ts
index 205e92b78f11..72fc8d6c501b 100644
--- a/clients/client-sagemaker/src/protocols/Aws_json1_1.ts
+++ b/clients/client-sagemaker/src/protocols/Aws_json1_1.ts
@@ -1374,7 +1374,6 @@ import {
ProductionVariantStatus,
ProductionVariantSummary,
ProfilerRuleEvaluationStatus,
- PropertyNameQuery,
RecommendationMetrics,
RStudioServerProDomainSettingsForUpdate,
SecondaryStatusTransition,
@@ -1575,6 +1574,7 @@ import {
ProfilerConfigForUpdate,
Project,
ProjectSummary,
+ PropertyNameQuery,
PutModelPackageGroupPolicyInput,
QueryFilters,
QueryLineageRequest,
@@ -1592,7 +1592,6 @@ import {
TransformJob,
TransformJobSummary,
Trial,
- TrialComponent,
TrialComponentSimpleSummary,
TrialComponentSourceDetail,
TrialComponentSummary,
@@ -1626,6 +1625,7 @@ import {
StopProcessingJobRequest,
StopTrainingJobRequest,
StopTransformJobRequest,
+ TrialComponent,
UpdateActionRequest,
UpdateAppImageConfigRequest,
UpdateArtifactRequest,
@@ -27473,6 +27473,7 @@ const de_RecommendationMetrics = (output: any, context: __SerdeContext): Recomme
MaxInvocations: __expectInt32,
MemoryUtilization: __limitedParseFloat32,
ModelLatency: __expectInt32,
+ ModelSetupTime: __expectInt32,
}) as any;
};
diff --git a/codegen/sdk-codegen/aws-models/sagemaker.json b/codegen/sdk-codegen/aws-models/sagemaker.json
index 0846fe87c9a5..5add38c26e28 100644
--- a/codegen/sdk-codegen/aws-models/sagemaker.json
+++ b/codegen/sdk-codegen/aws-models/sagemaker.json
@@ -3316,7 +3316,7 @@
"TimeSeriesForecastingJobConfig": {
"target": "com.amazonaws.sagemaker#TimeSeriesForecastingJobConfig",
"traits": {
- "smithy.api#documentation": "Settings used to configure an AutoML job V2 for a time-series forecasting problem\n type.
"
+ "smithy.api#documentation": "Settings used to configure an AutoML job V2 for a time-series forecasting problem\n type.
\n \n The TimeSeriesForecastingJobConfig problem type is only available in\n private beta. Contact Amazon Web Services Support or your account manager to learn more\n about access privileges.
\n "
}
}
},
@@ -22870,8 +22870,7 @@
"InstanceType": {
"target": "com.amazonaws.sagemaker#ProductionVariantInstanceType",
"traits": {
- "smithy.api#documentation": "The instance types to use for the load test.
",
- "smithy.api#required": {}
+ "smithy.api#documentation": "The instance types to use for the load test.
"
}
},
"InferenceSpecificationName": {
@@ -22885,6 +22884,9 @@
"traits": {
"smithy.api#documentation": " The parameter you want to benchmark against.
"
}
+ },
+ "ServerlessConfig": {
+ "target": "com.amazonaws.sagemaker#ProductionVariantServerlessConfig"
}
},
"traits": {
@@ -22976,17 +22978,17 @@
"InstanceType": {
"target": "com.amazonaws.sagemaker#ProductionVariantInstanceType",
"traits": {
- "smithy.api#documentation": "The instance type recommended by Amazon SageMaker Inference Recommender.
",
- "smithy.api#required": {}
+ "smithy.api#documentation": "The instance type recommended by Amazon SageMaker Inference Recommender.
"
}
},
"InitialInstanceCount": {
- "target": "com.amazonaws.sagemaker#Integer",
+ "target": "com.amazonaws.sagemaker#InitialInstanceCount",
"traits": {
- "smithy.api#default": 0,
- "smithy.api#documentation": "The number of instances recommended to launch initially.
",
- "smithy.api#required": {}
+ "smithy.api#documentation": "The number of instances recommended to launch initially.
"
}
+ },
+ "ServerlessConfig": {
+ "target": "com.amazonaws.sagemaker#ProductionVariantServerlessConfig"
}
},
"traits": {
@@ -28156,6 +28158,14 @@
"smithy.api#pattern": "^[a-zA-Z0-9](-*[a-zA-Z0-9]){0,62}$"
}
},
+ "com.amazonaws.sagemaker#InitialInstanceCount": {
+ "type": "integer",
+ "traits": {
+ "smithy.api#range": {
+ "min": 1
+ }
+ }
+ },
"com.amazonaws.sagemaker#InitialNumberOfUsers": {
"type": "integer",
"traits": {
@@ -39574,6 +39584,14 @@
"smithy.api#documentation": "The model registry settings for the SageMaker Canvas application.
"
}
},
+ "com.amazonaws.sagemaker#ModelSetupTime": {
+ "type": "integer",
+ "traits": {
+ "smithy.api#range": {
+ "min": 0
+ }
+ }
+ },
"com.amazonaws.sagemaker#ModelSortKey": {
"type": "enum",
"members": {
@@ -45284,7 +45302,7 @@
"ProvisionedConcurrency": {
"target": "com.amazonaws.sagemaker#ServerlessProvisionedConcurrency",
"traits": {
- "smithy.api#documentation": "The amount of provisioned concurrency to allocate for the serverless endpoint.\n Should be less than or equal to MaxConcurrency.
"
+ "smithy.api#documentation": "The amount of provisioned concurrency to allocate for the serverless endpoint.\n Should be less than or equal to MaxConcurrency.
\n \n This field is not supported for serverless endpoint recommendations for Inference Recommender jobs.\n For more information about creating an Inference Recommender job, see\n CreateInferenceRecommendationsJobs.
\n "
}
}
},
@@ -46655,6 +46673,12 @@
"traits": {
"smithy.api#documentation": "Specifies the name and shape of the expected data inputs for your trained model with a JSON dictionary form.\n This field is used for optimizing your model using SageMaker Neo. For more information, see\n DataInputConfig.
"
}
+ },
+ "SupportedEndpointType": {
+ "target": "com.amazonaws.sagemaker#RecommendationJobSupportedEndpointType",
+ "traits": {
+ "smithy.api#documentation": "The endpoint type to receive recommendations for. By default this is null, and the results of \n the inference recommendation job return a combined list of both real-time and serverless benchmarks.\n By specifying a value for this field, you can receive a longer list of benchmarks for the desired endpoint type.
"
+ }
}
},
"traits": {
@@ -46927,6 +46951,23 @@
"target": "com.amazonaws.sagemaker#String"
}
},
+ "com.amazonaws.sagemaker#RecommendationJobSupportedEndpointType": {
+ "type": "enum",
+ "members": {
+ "REALTIME": {
+ "target": "smithy.api#Unit",
+ "traits": {
+ "smithy.api#enumValue": "RealTime"
+ }
+ },
+ "SERVERLESS": {
+ "target": "smithy.api#Unit",
+ "traits": {
+ "smithy.api#enumValue": "Serverless"
+ }
+ }
+ }
+ },
"com.amazonaws.sagemaker#RecommendationJobSupportedInstanceTypes": {
"type": "list",
"member": {
@@ -47062,6 +47103,12 @@
"traits": {
"smithy.api#documentation": "The expected memory utilization at maximum invocations per minute for the instance.
\n \n NaN indicates that the value is not available.
"
}
+ },
+ "ModelSetupTime": {
+ "target": "com.amazonaws.sagemaker#ModelSetupTime",
+ "traits": {
+ "smithy.api#documentation": "The time it takes to launch new compute resources for a serverless endpoint.\n The time can vary depending on the model size, how long it takes to download the\n model, and the start-up time of the container.
\n \n NaN indicates that the value is not available.
"
+ }
}
},
"traits": {
@@ -47990,7 +48037,6 @@
"MaximumBatchSize": {
"target": "com.amazonaws.sagemaker#CapacitySize",
"traits": {
- "smithy.api#documentation": "Batch size for each rolling step to provision capacity and turn on traffic on the new\n endpoint fleet, and terminate capacity on the old endpoint fleet. Value must be between\n 5% to 50% of the variant's total instance count.
",
"smithy.api#required": {}
}
},
@@ -48008,10 +48054,7 @@
}
},
"RollbackMaximumBatchSize": {
- "target": "com.amazonaws.sagemaker#CapacitySize",
- "traits": {
- "smithy.api#documentation": "Batch size for rollback to the old endpoint fleet. Each rolling step to provision\n capacity and turn on traffic on the old endpoint fleet, and terminate capacity on the new\n endpoint fleet. If this field is absent, the default value will be set to 100% of total\n capacity which means to bring up the whole capacity of the old fleet at once during rollback.
"
- }
+ "target": "com.amazonaws.sagemaker#CapacitySize"
}
},
"traits": {