Skip to content

Commit 8680432

Browse files
author
awstools
committed
feat(client-sagemaker): Enhanced SageMaker HyperPod instance groups with support for MinInstanceCount, CapacityRequirements (Spot/On-Demand), and KubernetesConfig (labels and taints). Also Added speculative decoding and MaxInstanceCount for model optimization jobs.
1 parent 1fb4616 commit 8680432

28 files changed

+6202
-5084
lines changed

‎clients/client-sagemaker/src/commands/CreateClusterCommand.ts‎

Lines changed: 17 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -41,6 +41,7 @@ export interface CreateClusterCommandOutput extends CreateClusterResponse, __Met
4141
* InstanceGroups: [ // ClusterInstanceGroupSpecifications
4242
* { // ClusterInstanceGroupSpecification
4343
* InstanceCount: Number("int"), // required
44+
* MinInstanceCount: Number("int"),
4445
* InstanceGroupName: "STRING_VALUE", // required
4546
* InstanceType: "ml.p4d.24xlarge" || "ml.p4de.24xlarge" || "ml.p5.48xlarge" || "ml.p6e-gb200.36xlarge" || "ml.trn1.32xlarge" || "ml.trn1n.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.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.12xlarge" || "ml.c5.18xlarge" || "ml.c5.24xlarge" || "ml.c5n.large" || "ml.c5n.2xlarge" || "ml.c5n.4xlarge" || "ml.c5n.9xlarge" || "ml.c5n.18xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.8xlarge" || "ml.m5.12xlarge" || "ml.m5.16xlarge" || "ml.m5.24xlarge" || "ml.t3.medium" || "ml.t3.large" || "ml.t3.xlarge" || "ml.t3.2xlarge" || "ml.g6.xlarge" || "ml.g6.2xlarge" || "ml.g6.4xlarge" || "ml.g6.8xlarge" || "ml.g6.16xlarge" || "ml.g6.12xlarge" || "ml.g6.24xlarge" || "ml.g6.48xlarge" || "ml.gr6.4xlarge" || "ml.gr6.8xlarge" || "ml.g6e.xlarge" || "ml.g6e.2xlarge" || "ml.g6e.4xlarge" || "ml.g6e.8xlarge" || "ml.g6e.16xlarge" || "ml.g6e.12xlarge" || "ml.g6e.24xlarge" || "ml.g6e.48xlarge" || "ml.p5e.48xlarge" || "ml.p5en.48xlarge" || "ml.p6-b200.48xlarge" || "ml.trn2.3xlarge" || "ml.trn2.48xlarge" || "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.m6i.large" || "ml.m6i.xlarge" || "ml.m6i.2xlarge" || "ml.m6i.4xlarge" || "ml.m6i.8xlarge" || "ml.m6i.12xlarge" || "ml.m6i.16xlarge" || "ml.m6i.24xlarge" || "ml.m6i.32xlarge" || "ml.r6i.large" || "ml.r6i.xlarge" || "ml.r6i.2xlarge" || "ml.r6i.4xlarge" || "ml.r6i.8xlarge" || "ml.r6i.12xlarge" || "ml.r6i.16xlarge" || "ml.r6i.24xlarge" || "ml.r6i.32xlarge" || "ml.i3en.large" || "ml.i3en.xlarge" || "ml.i3en.2xlarge" || "ml.i3en.3xlarge" || "ml.i3en.6xlarge" || "ml.i3en.12xlarge" || "ml.i3en.24xlarge" || "ml.m7i.large" || "ml.m7i.xlarge" || "ml.m7i.2xlarge" || "ml.m7i.4xlarge" || "ml.m7i.8xlarge" || "ml.m7i.12xlarge" || "ml.m7i.16xlarge" || "ml.m7i.24xlarge" || "ml.m7i.48xlarge" || "ml.r7i.large" || "ml.r7i.xlarge" || "ml.r7i.2xlarge" || "ml.r7i.4xlarge" || "ml.r7i.8xlarge" || "ml.r7i.12xlarge" || "ml.r7i.16xlarge" || "ml.r7i.24xlarge" || "ml.r7i.48xlarge", // required
4647
* LifeCycleConfig: { // ClusterLifeCycleConfig
@@ -92,6 +93,22 @@ export interface CreateClusterCommandOutput extends CreateClusterResponse, __Met
9293
* },
9394
* },
9495
* ImageId: "STRING_VALUE",
96+
* KubernetesConfig: { // ClusterKubernetesConfig
97+
* Labels: { // ClusterKubernetesLabels
98+
* "<keys>": "STRING_VALUE",
99+
* },
100+
* Taints: [ // ClusterKubernetesTaints
101+
* { // ClusterKubernetesTaint
102+
* Key: "STRING_VALUE", // required
103+
* Value: "STRING_VALUE",
104+
* Effect: "NoSchedule" || "PreferNoSchedule" || "NoExecute", // required
105+
* },
106+
* ],
107+
* },
108+
* CapacityRequirements: { // ClusterCapacityRequirements
109+
* Spot: {},
110+
* OnDemand: {},
111+
* },
95112
* },
96113
* ],
97114
* RestrictedInstanceGroups: [ // ClusterRestrictedInstanceGroupSpecifications

‎clients/client-sagemaker/src/commands/CreateOptimizationJobCommand.ts‎

Lines changed: 15 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -46,8 +46,12 @@ export interface CreateOptimizationJobCommandOutput extends CreateOptimizationJo
4646
* AcceptEula: true || false, // required
4747
* },
4848
* },
49+
* SageMakerModel: { // OptimizationSageMakerModel
50+
* ModelName: "STRING_VALUE",
51+
* },
4952
* },
50-
* DeploymentInstanceType: "ml.p4d.24xlarge" || "ml.p4de.24xlarge" || "ml.p5.48xlarge" || "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.g6.xlarge" || "ml.g6.2xlarge" || "ml.g6.4xlarge" || "ml.g6.8xlarge" || "ml.g6.12xlarge" || "ml.g6.16xlarge" || "ml.g6.24xlarge" || "ml.g6.48xlarge" || "ml.g6e.xlarge" || "ml.g6e.2xlarge" || "ml.g6e.4xlarge" || "ml.g6e.8xlarge" || "ml.g6e.12xlarge" || "ml.g6e.16xlarge" || "ml.g6e.24xlarge" || "ml.g6e.48xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.trn1n.32xlarge", // required
53+
* DeploymentInstanceType: "ml.p4d.24xlarge" || "ml.p4de.24xlarge" || "ml.p5.48xlarge" || "ml.p5e.48xlarge" || "ml.p5en.48xlarge" || "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.g6.xlarge" || "ml.g6.2xlarge" || "ml.g6.4xlarge" || "ml.g6.8xlarge" || "ml.g6.12xlarge" || "ml.g6.16xlarge" || "ml.g6.24xlarge" || "ml.g6.48xlarge" || "ml.g6e.xlarge" || "ml.g6e.2xlarge" || "ml.g6e.4xlarge" || "ml.g6e.8xlarge" || "ml.g6e.12xlarge" || "ml.g6e.16xlarge" || "ml.g6e.24xlarge" || "ml.g6e.48xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.trn1n.32xlarge", // required
54+
* MaxInstanceCount: Number("int"),
5155
* OptimizationEnvironment: { // OptimizationJobEnvironmentVariables
5256
* "<keys>": "STRING_VALUE",
5357
* },
@@ -71,11 +75,21 @@ export interface CreateOptimizationJobCommandOutput extends CreateOptimizationJo
7175
* "<keys>": "STRING_VALUE",
7276
* },
7377
* },
78+
* ModelSpeculativeDecodingConfig: { // ModelSpeculativeDecodingConfig
79+
* Technique: "EAGLE", // required
80+
* TrainingDataSource: { // ModelSpeculativeDecodingTrainingDataSource
81+
* S3Uri: "STRING_VALUE", // required
82+
* S3DataType: "S3Prefix" || "ManifestFile", // required
83+
* },
84+
* },
7485
* },
7586
* ],
7687
* OutputConfig: { // OptimizationJobOutputConfig
7788
* KmsKeyId: "STRING_VALUE",
7889
* S3OutputLocation: "STRING_VALUE", // required
90+
* SageMakerModel: {
91+
* ModelName: "STRING_VALUE",
92+
* },
7993
* },
8094
* StoppingCondition: { // StoppingCondition
8195
* MaxRuntimeInSeconds: Number("int"),

‎clients/client-sagemaker/src/commands/CreateTrainingJobCommand.ts‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,8 @@ import { Command as $Command } from "@smithy/smithy-client";
44
import { MetadataBearer as __MetadataBearer } from "@smithy/types";
55

66
import { commonParams } from "../endpoint/EndpointParameters";
7-
import { CreateTrainingJobRequest, CreateTrainingJobResponse } from "../models/models_1";
7+
import { CreateTrainingJobRequest } from "../models/models_1";
8+
import { CreateTrainingJobResponse } from "../models/models_2";
89
import { SageMakerClientResolvedConfig, ServiceInputTypes, ServiceOutputTypes } from "../SageMakerClient";
910
import { CreateTrainingJob } from "../schemas/schemas_0";
1011

‎clients/client-sagemaker/src/commands/CreateTrainingPlanCommand.ts‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ import { Command as $Command } from "@smithy/smithy-client";
44
import { MetadataBearer as __MetadataBearer } from "@smithy/types";
55

66
import { commonParams } from "../endpoint/EndpointParameters";
7-
import { CreateTrainingPlanRequest, CreateTrainingPlanResponse } from "../models/models_1";
7+
import { CreateTrainingPlanRequest, CreateTrainingPlanResponse } from "../models/models_2";
88
import { SageMakerClientResolvedConfig, ServiceInputTypes, ServiceOutputTypes } from "../SageMakerClient";
99
import { CreateTrainingPlan } from "../schemas/schemas_0";
1010

‎clients/client-sagemaker/src/commands/CreateTransformJobCommand.ts‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ import { Command as $Command } from "@smithy/smithy-client";
44
import { MetadataBearer as __MetadataBearer } from "@smithy/types";
55

66
import { commonParams } from "../endpoint/EndpointParameters";
7-
import { CreateTransformJobRequest, CreateTransformJobResponse } from "../models/models_1";
7+
import { CreateTransformJobRequest, CreateTransformJobResponse } from "../models/models_2";
88
import { SageMakerClientResolvedConfig, ServiceInputTypes, ServiceOutputTypes } from "../SageMakerClient";
99
import { CreateTransformJob } from "../schemas/schemas_0";
1010

‎clients/client-sagemaker/src/commands/CreateTrialCommand.ts‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ import { Command as $Command } from "@smithy/smithy-client";
44
import { MetadataBearer as __MetadataBearer } from "@smithy/types";
55

66
import { commonParams } from "../endpoint/EndpointParameters";
7-
import { CreateTrialRequest, CreateTrialResponse } from "../models/models_1";
7+
import { CreateTrialRequest, CreateTrialResponse } from "../models/models_2";
88
import { SageMakerClientResolvedConfig, ServiceInputTypes, ServiceOutputTypes } from "../SageMakerClient";
99
import { CreateTrial } from "../schemas/schemas_0";
1010

‎clients/client-sagemaker/src/commands/DescribeClusterCommand.ts‎

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -51,6 +51,7 @@ export interface DescribeClusterCommandOutput extends DescribeClusterResponse, _
5151
* // { // ClusterInstanceGroupDetails
5252
* // CurrentCount: Number("int"),
5353
* // TargetCount: Number("int"),
54+
* // MinCount: Number("int"),
5455
* // InstanceGroupName: "STRING_VALUE",
5556
* // InstanceType: "ml.p4d.24xlarge" || "ml.p4de.24xlarge" || "ml.p5.48xlarge" || "ml.p6e-gb200.36xlarge" || "ml.trn1.32xlarge" || "ml.trn1n.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.c5.large" || "ml.c5.xlarge" || "ml.c5.2xlarge" || "ml.c5.4xlarge" || "ml.c5.9xlarge" || "ml.c5.12xlarge" || "ml.c5.18xlarge" || "ml.c5.24xlarge" || "ml.c5n.large" || "ml.c5n.2xlarge" || "ml.c5n.4xlarge" || "ml.c5n.9xlarge" || "ml.c5n.18xlarge" || "ml.m5.large" || "ml.m5.xlarge" || "ml.m5.2xlarge" || "ml.m5.4xlarge" || "ml.m5.8xlarge" || "ml.m5.12xlarge" || "ml.m5.16xlarge" || "ml.m5.24xlarge" || "ml.t3.medium" || "ml.t3.large" || "ml.t3.xlarge" || "ml.t3.2xlarge" || "ml.g6.xlarge" || "ml.g6.2xlarge" || "ml.g6.4xlarge" || "ml.g6.8xlarge" || "ml.g6.16xlarge" || "ml.g6.12xlarge" || "ml.g6.24xlarge" || "ml.g6.48xlarge" || "ml.gr6.4xlarge" || "ml.gr6.8xlarge" || "ml.g6e.xlarge" || "ml.g6e.2xlarge" || "ml.g6e.4xlarge" || "ml.g6e.8xlarge" || "ml.g6e.16xlarge" || "ml.g6e.12xlarge" || "ml.g6e.24xlarge" || "ml.g6e.48xlarge" || "ml.p5e.48xlarge" || "ml.p5en.48xlarge" || "ml.p6-b200.48xlarge" || "ml.trn2.3xlarge" || "ml.trn2.48xlarge" || "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.m6i.large" || "ml.m6i.xlarge" || "ml.m6i.2xlarge" || "ml.m6i.4xlarge" || "ml.m6i.8xlarge" || "ml.m6i.12xlarge" || "ml.m6i.16xlarge" || "ml.m6i.24xlarge" || "ml.m6i.32xlarge" || "ml.r6i.large" || "ml.r6i.xlarge" || "ml.r6i.2xlarge" || "ml.r6i.4xlarge" || "ml.r6i.8xlarge" || "ml.r6i.12xlarge" || "ml.r6i.16xlarge" || "ml.r6i.24xlarge" || "ml.r6i.32xlarge" || "ml.i3en.large" || "ml.i3en.xlarge" || "ml.i3en.2xlarge" || "ml.i3en.3xlarge" || "ml.i3en.6xlarge" || "ml.i3en.12xlarge" || "ml.i3en.24xlarge" || "ml.m7i.large" || "ml.m7i.xlarge" || "ml.m7i.2xlarge" || "ml.m7i.4xlarge" || "ml.m7i.8xlarge" || "ml.m7i.12xlarge" || "ml.m7i.16xlarge" || "ml.m7i.24xlarge" || "ml.m7i.48xlarge" || "ml.r7i.large" || "ml.r7i.xlarge" || "ml.r7i.2xlarge" || "ml.r7i.4xlarge" || "ml.r7i.8xlarge" || "ml.r7i.12xlarge" || "ml.r7i.16xlarge" || "ml.r7i.24xlarge" || "ml.r7i.48xlarge",
5657
* // LifeCycleConfig: { // ClusterLifeCycleConfig
@@ -105,6 +106,35 @@ export interface DescribeClusterCommandOutput extends DescribeClusterResponse, _
105106
* // },
106107
* // CurrentImageId: "STRING_VALUE",
107108
* // DesiredImageId: "STRING_VALUE",
109+
* // ActiveOperations: { // ActiveOperations
110+
* // "<keys>": Number("int"),
111+
* // },
112+
* // KubernetesConfig: { // ClusterKubernetesConfigDetails
113+
* // CurrentLabels: { // ClusterKubernetesLabels
114+
* // "<keys>": "STRING_VALUE",
115+
* // },
116+
* // DesiredLabels: {
117+
* // "<keys>": "STRING_VALUE",
118+
* // },
119+
* // CurrentTaints: [ // ClusterKubernetesTaints
120+
* // { // ClusterKubernetesTaint
121+
* // Key: "STRING_VALUE", // required
122+
* // Value: "STRING_VALUE",
123+
* // Effect: "NoSchedule" || "PreferNoSchedule" || "NoExecute", // required
124+
* // },
125+
* // ],
126+
* // DesiredTaints: [
127+
* // {
128+
* // Key: "STRING_VALUE", // required
129+
* // Value: "STRING_VALUE",
130+
* // Effect: "NoSchedule" || "PreferNoSchedule" || "NoExecute", // required
131+
* // },
132+
* // ],
133+
* // },
134+
* // CapacityRequirements: { // ClusterCapacityRequirements
135+
* // Spot: {},
136+
* // OnDemand: {},
137+
* // },
108138
* // TargetStateCount: Number("int"),
109139
* // SoftwareUpdateStatus: "Pending" || "InProgress" || "Succeeded" || "Failed" || "RollbackInProgress" || "RollbackComplete",
110140
* // ActiveSoftwareUpdateConfig: {

‎clients/client-sagemaker/src/commands/DescribeClusterEventCommand.ts‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -76,6 +76,7 @@ export interface DescribeClusterEventCommandOutput extends DescribeClusterEventR
7676
* // InstanceGroupScaling: { // InstanceGroupScalingMetadata
7777
* // InstanceCount: Number("int"),
7878
* // TargetCount: Number("int"),
79+
* // MinCount: Number("int"),
7980
* // FailureMessage: "STRING_VALUE",
8081
* // },
8182
* // Instance: { // InstanceMetadata

‎clients/client-sagemaker/src/commands/DescribeClusterNodeCommand.ts‎

Lines changed: 23 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -89,6 +89,29 @@ export interface DescribeClusterNodeCommandOutput extends DescribeClusterNodeRes
8989
* // UltraServerInfo: { // UltraServerInfo
9090
* // Id: "STRING_VALUE",
9191
* // },
92+
* // KubernetesConfig: { // ClusterKubernetesConfigNodeDetails
93+
* // CurrentLabels: { // ClusterKubernetesLabels
94+
* // "<keys>": "STRING_VALUE",
95+
* // },
96+
* // DesiredLabels: {
97+
* // "<keys>": "STRING_VALUE",
98+
* // },
99+
* // CurrentTaints: [ // ClusterKubernetesTaints
100+
* // { // ClusterKubernetesTaint
101+
* // Key: "STRING_VALUE", // required
102+
* // Value: "STRING_VALUE",
103+
* // Effect: "NoSchedule" || "PreferNoSchedule" || "NoExecute", // required
104+
* // },
105+
* // ],
106+
* // DesiredTaints: [
107+
* // {
108+
* // Key: "STRING_VALUE", // required
109+
* // Value: "STRING_VALUE",
110+
* // Effect: "NoSchedule" || "PreferNoSchedule" || "NoExecute", // required
111+
* // },
112+
* // ],
113+
* // },
114+
* // CapacityType: "Spot" || "OnDemand",
92115
* // },
93116
* // };
94117
*

‎clients/client-sagemaker/src/commands/DescribeOptimizationJobCommand.ts‎

Lines changed: 15 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -57,11 +57,15 @@ export interface DescribeOptimizationJobCommandOutput extends DescribeOptimizati
5757
* // AcceptEula: true || false, // required
5858
* // },
5959
* // },
60+
* // SageMakerModel: { // OptimizationSageMakerModel
61+
* // ModelName: "STRING_VALUE",
62+
* // },
6063
* // },
6164
* // OptimizationEnvironment: { // OptimizationJobEnvironmentVariables
6265
* // "<keys>": "STRING_VALUE",
6366
* // },
64-
* // DeploymentInstanceType: "ml.p4d.24xlarge" || "ml.p4de.24xlarge" || "ml.p5.48xlarge" || "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.g6.xlarge" || "ml.g6.2xlarge" || "ml.g6.4xlarge" || "ml.g6.8xlarge" || "ml.g6.12xlarge" || "ml.g6.16xlarge" || "ml.g6.24xlarge" || "ml.g6.48xlarge" || "ml.g6e.xlarge" || "ml.g6e.2xlarge" || "ml.g6e.4xlarge" || "ml.g6e.8xlarge" || "ml.g6e.12xlarge" || "ml.g6e.16xlarge" || "ml.g6e.24xlarge" || "ml.g6e.48xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.trn1n.32xlarge", // required
67+
* // DeploymentInstanceType: "ml.p4d.24xlarge" || "ml.p4de.24xlarge" || "ml.p5.48xlarge" || "ml.p5e.48xlarge" || "ml.p5en.48xlarge" || "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.g6.xlarge" || "ml.g6.2xlarge" || "ml.g6.4xlarge" || "ml.g6.8xlarge" || "ml.g6.12xlarge" || "ml.g6.16xlarge" || "ml.g6.24xlarge" || "ml.g6.48xlarge" || "ml.g6e.xlarge" || "ml.g6e.2xlarge" || "ml.g6e.4xlarge" || "ml.g6e.8xlarge" || "ml.g6e.12xlarge" || "ml.g6e.16xlarge" || "ml.g6e.24xlarge" || "ml.g6e.48xlarge" || "ml.inf2.xlarge" || "ml.inf2.8xlarge" || "ml.inf2.24xlarge" || "ml.inf2.48xlarge" || "ml.trn1.2xlarge" || "ml.trn1.32xlarge" || "ml.trn1n.32xlarge", // required
68+
* // MaxInstanceCount: Number("int"),
6569
* // OptimizationConfigs: [ // OptimizationConfigs // required
6670
* // { // OptimizationConfig Union: only one key present
6771
* // ModelQuantizationConfig: { // ModelQuantizationConfig
@@ -82,11 +86,21 @@ export interface DescribeOptimizationJobCommandOutput extends DescribeOptimizati
8286
* // "<keys>": "STRING_VALUE",
8387
* // },
8488
* // },
89+
* // ModelSpeculativeDecodingConfig: { // ModelSpeculativeDecodingConfig
90+
* // Technique: "EAGLE", // required
91+
* // TrainingDataSource: { // ModelSpeculativeDecodingTrainingDataSource
92+
* // S3Uri: "STRING_VALUE", // required
93+
* // S3DataType: "S3Prefix" || "ManifestFile", // required
94+
* // },
95+
* // },
8596
* // },
8697
* // ],
8798
* // OutputConfig: { // OptimizationJobOutputConfig
8899
* // KmsKeyId: "STRING_VALUE",
89100
* // S3OutputLocation: "STRING_VALUE", // required
101+
* // SageMakerModel: {
102+
* // ModelName: "STRING_VALUE",
103+
* // },
90104
* // },
91105
* // OptimizationOutput: { // OptimizationOutput
92106
* // RecommendedInferenceImage: "STRING_VALUE",

0 commit comments

Comments
 (0)