AsyncInferenceConfig takes different parameters

Question

I would like to deploy an async endpoint in SageMaker. However when trying to deploy it I get the following error: 
ParamValidationError: Parameter validation failed:
Unknown parameter in input: "AsyncInferenceConfig"

This is the code I tried for deploying the endpoint
```
from sagemaker.async_inference.async_inference_config import AsyncInferenceConfig

async_config = AsyncInferenceConfig(
    output_path="s3://poembucketus/async_inference/output",
    max_concurrent_invocations_per_instance=4,
)

async_predictor = huggingface_estimator.deploy(
    initial_instance_count=1,
    instance_type="ml.m5.xlarge",
    async_inference_config=async_config,
)
```
Thanks!

Answer

From the format of parameter, this error should be thrown from Boto3. If AsyncInferenceConfig is not recognized as a valid parameter, this may mean the version of boto3 is too old and released before the async feature.

Could you please confirm the version of boto3 and try updating it to the latest if possible ?

AsyncInferenceConfig takes different parameters

Conteúdo relevante