config async problem when creating endpoint #4275
Labels
bug
This issue is a confirmed bug.
p2
This is a standard priority issue
response-requested
Waiting on additional information or feedback.
sagemaker
service-api
This issue is caused by the service API, not the SDK implementation.
Describe the bug
when i remove the async 'AsyncInferenceConfig' , it works, but i want an async endpoint because my payload is above 5 mb, its a numpy array of image of 50 + mb , so i need async, the container is triton
Expected Behavior
create an endpoint
Current Behavior
botocore.exceptions.ClientError: An error occurred (ValidationException) when calling the CreateEndpoint operation: One or more endpoint features are not supported using this configuration
Reproduction Steps
create_model = sagemaker_client.create_model(
ModelName = 'tritoninferencesasync',
ExecutionRoleArn = role,
PrimaryContainer= container)
create_endpoint_config_response = sm_client.create_endpoint_config(
EndpointConfigName='endpointconfigsingleasnyc',
ProductionVariants=[
{
"VariantName": "variant1",
"ModelName": 'tritoninferencesasync',
"InstanceType": "ml.m5.xlarge",
"InitialInstanceCount": 1,
}
],
AsyncInferenceConfig={
"OutputConfig": {
"S3OutputPath": "s3://allinferences/yy/xx"
},
"ClientConfig": {"MaxConcurrentInvocationsPerInstance": 4},
},
)
endpoint_name='tritonendpointsasync'
create_multi_endpoint = sagemaker_client.create_endpoint(
EndpointName = endpoint_name,
EndpointConfigName = 'endpointconfigsingleasnyc')
Possible Solution
another region maybe >? > currently in milan
Additional Information/Context
No response
SDK version used
boto3==1.35.11 botocore==1.35.11
Environment details (OS name and version, etc.)
windows 11
The text was updated successfully, but these errors were encountered: