Invalid base64: "{"instances": [{"in0":[863],"in1":[882]}]}" when testing Amazon SageMaker model endpoint using the AWS CLI

Question

I am new to Amazon SageMaker and I am closely following this tutorial https://aws.amazon.com/blogs/machine-learning/creating-a-machine-learning-powered-rest-api-with-amazon-api-gateway-mapping-templates-and-amazon-sagemaker/ to create a machine learning-powered REST API with Amazon API Gateway mapping templates and Amazon SageMaker

when I run the following command on terminal (Step 2 of the Tutorial )

aws sagemaker-runtime invoke-endpoint \
  --endpoint-name <endpoint-name> \
  --body '{"instances": [{"in0":[863],"in1":[882]}]}' \
  --content-type application/json \
  --accept application/json \
  results

I get the following Error: Invalid base64: "{"instances": [{"in0":[863],"in1":[882]}]}" My endpoint is InService on the SageMaker console and the example Jupyter notebook run successfully. (I also substituted <endpoint-name> with the actual name - same error received with/without quotations around the name)

Using zsh here is the aws cli version:

aws --version
aws-cli/2.0.15 Python/3.7.4 Darwin/19.4.0 botocore/2.0.0dev19

Wondering what the problem could be. Any help is appreciated

Answer 1

The problem is that the body contents is being expected to be base 64 encoded, try base64 encoding the body before passing it to the invoke statement.