Amazon SageMaker Local Mode raised boto3.exceptions.RetriesExceededError: Max Retries Exceeded

Question

I try to run the SageMaker local mode example without any modification at https://github.com/aws-samples/amazon-sagemaker-local-mode/tree/main/pytorch_nlp_script_mode_local_model_inference on my local machine.

However I encountered the **boto3.exceptions.RetriesExceededError: Max Retries Exceeded
** exception when the example tries to deploy the inference endpoint to 'local' instance type.

I checked with 
```
docker images -a 
```
command and it does not pull the expected pre-built SageMaker deep learning container image from ECR. The code example is using a dummy role for the local SageMaker session. I need help as I am blocked at this point as the exception error message is not helpful to pinpoint the actual root cause of this issue. Thanks in advance.

Below are my configurations:

* Ubuntu: **20.04.3 LTS**
* AWS CLI version: **2.4.7**
* Python: **3.8.12**
* Docker: **20.10.12**
* Docker Compose: **1.29.2**
* boto3: **1.20.26**
* sagemaker: **2.72.1**

Answer

Hi, if you aren't able to pull the expected pre-built SageMaker deep learning container image, I would check your network settings.  From the local mode sample documentation, "you'll need to be able to access a public Amazon ECR repository from your local environment."

For more detailed information from boto3, you can enable logging and debug mode as well, which could help pinpoint the exact cause of the error.  More information on Boto3 Retries: https://boto3.amazonaws.com/v1/documentation/api/latest/guide/retries.html.

Amazon SageMaker Local Mode raised boto3.exceptions.RetriesExceededError: Max Retries Exceeded

관련 콘텐츠