Slight fixes to sagemaker inference sample

Author: silv-io

sagemaker-inference/ReadMe.md

@@ -5,5 +5,7 @@ This is a small example about how you can use LocalStack to host your PyTorch ML
 Before using this example you should setup your Docker Client to pull the AWS Deep Learning images ([more info here](https://github.com/aws/deep-learning-containers/blob/master/available_images.md)):
 
 ```bash
-    aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin 763104351884.dkr.ecr.us-east-1.amazonaws.com
+    aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin 763104351884.dkr.ecr.us-east-1.amazonaws.com
 ```
+
+Because the images tend to be really big (multiple GB), you might want to `docker pull` them beforehand to avoid any timeouts.

sagemaker-inference/main.py

@@ -20,9 +20,10 @@
 CONTAINER_IMAGE = "763104351884.dkr.ecr.us-east-1.amazonaws.com/pytorch-inference:1.5.0-cpu-py3"
 EXECUTION_ROLE_ARN = "arn:aws:iam::0000000000000:role/sagemaker-role"
 
-sagemaker: SageMakerClient = boto3.client("sagemaker", endpoint_url=LOCALSTACK_ENDPOINT)
-sagemaker_runtime: SageMakerRuntimeClient = boto3.client("sagemaker-runtime", endpoint_url=LOCALSTACK_ENDPOINT)
-s3: S3Client = boto3.client("s3", endpoint_url=LOCALSTACK_ENDPOINT)
+sagemaker: SageMakerClient = boto3.client("sagemaker", endpoint_url=LOCALSTACK_ENDPOINT, region_name="us-east-1")
+sagemaker_runtime: SageMakerRuntimeClient = boto3.client("sagemaker-runtime", endpoint_url=LOCALSTACK_ENDPOINT,
+                                                         region_name="us-east-1")
+s3: S3Client = boto3.client("s3", endpoint_url=LOCALSTACK_ENDPOINT, region_name="us-east-1")
 
 
 def deploy_model(run_id: str = "0"):