interface InferenceComponentSpecificationForInstanceTypeProperty
| Language | Type name |
|---|---|
 .NET | Amazon.CDK.AWS.Sagemaker.CfnInferenceComponent.InferenceComponentSpecificationForInstanceTypeProperty |
 Go | github.com/aws/aws-cdk-go/awscdk/v2/awssagemaker#CfnInferenceComponent_InferenceComponentSpecificationForInstanceTypeProperty |
 Java | software.amazon.awscdk.services.sagemaker.CfnInferenceComponent.InferenceComponentSpecificationForInstanceTypeProperty |
 Python | aws_cdk.aws_sagemaker.CfnInferenceComponent.InferenceComponentSpecificationForInstanceTypeProperty |
 TypeScript | aws-cdk-lib » aws_sagemaker » CfnInferenceComponent » InferenceComponentSpecificationForInstanceTypeProperty |
A specification for one instance type, for use in Specifications.
InstanceType is required here, and is not accepted on the singular Specification. BaseInferenceComponentName is not accepted here either: adapter inference components are supported only on the singular Specification.
Example
// The code below shows an example of how to instantiate this type.
// The values are placeholders you should change.
import { aws_sagemaker as sagemaker } from 'aws-cdk-lib';
const inferenceComponentSpecificationForInstanceTypeProperty: sagemaker.CfnInferenceComponent.InferenceComponentSpecificationForInstanceTypeProperty = {
instanceType: 'instanceType',
// the properties below are optional
computeResourceRequirements: {
maxMemoryRequiredInMb: 123,
minMemoryRequiredInMb: 123,
numberOfAcceleratorDevicesRequired: 123,
numberOfCpuCoresRequired: 123,
},
container: {
artifactUrl: 'artifactUrl',
containerMetricsConfig: {
metricsEndpoints: [{
metricsEndpointPath: 'metricsEndpointPath',
// the properties below are optional
metricPublishFrequencyInSeconds: 123,
}],
},
environment: {
environmentKey: 'environment',
},
image: 'image',
},
currentDataCacheConfig: {
enableCaching: false,
},
dataCacheConfig: {
enableCaching: false,
},
modelName: 'modelName',
schedulingConfig: {
availabilityZoneBalance: {
enforcementMode: 'enforcementMode',
// the properties below are optional
maxImbalance: 123,
},
placementStrategy: 'placementStrategy',
},
startupParameters: {
containerStartupHealthCheckTimeoutInSeconds: 123,
modelDataDownloadTimeoutInSeconds: 123,
},
};
Properties
| Name | Type | Description |
|---|---|---|
| instance | string | An ML compute instance type. |
| compute | IResolvable | Inference | |
| container? | IResolvable | Inference | Container specification for one Specifications entry. |
| current | IResolvable | Inference | Settings that affect how the inference component caches data. |
| data | IResolvable | Inference | Settings that affect how the inference component caches data. |
| model | string | The name of the model to use with the inference component. |
| scheduling | IResolvable | Inference | The scheduling configuration that determines how inference component copies are placed across available instances. |
| startup | IResolvable | Inference |
instanceType
Type:
string
An ML compute instance type.
computeResourceRequirements?
Type:
IResolvable | Inference
(optional)
container?
Type:
IResolvable | Inference
(optional)
Container specification for one Specifications entry.
Distinct from InferenceComponentContainerSpecification: DescribeInferenceComponent returns no per-entry DeployedImage (VERIFIED in us-west-2), so DeployedImage is intentionally omitted here and this definition can never be aggregated into a plural READ response. The singular InferenceComponentContainerSpecification keeps DeployedImage - the service DOES return it there.
currentDataCacheConfig?
Type:
IResolvable | Inference
(optional)
Settings that affect how the inference component caches data.
dataCacheConfig?
Type:
IResolvable | Inference
(optional)
Settings that affect how the inference component caches data.
modelName?
Type:
string
(optional)
The name of the model to use with the inference component.
schedulingConfig?
Type:
IResolvable | Inference
(optional)
The scheduling configuration that determines how inference component copies are placed across available instances.
startupParameters?
Type:
IResolvable | Inference
(optional)