如何部署 sagemaker 模型（使用 Java）

Question

I am trying to use Sagemaker for predictions that my Springboot application needs. I have created an ML model using sageMakerClient.createModel then created a training job for it using sageMakerClient.createTrainingJob and successfully created model as shown below,

Now how to I deploy the model and create a REST end point that I can use from my Java application to end data and get predictions? I am not able to figure that out.

Answer 1

You should be able to do this using the Java SDK. Here is the documentation for creating an endpoint. http://appdoc.app/artifact/com.amazonaws/aws-java-sdk-osgi/1.11.333/com/amazonaws/services/sagemaker/model/CreateEndpointResult.html

To invoke an endpoint use the following documentation https://docs.aws.amazon.com/AWSJavaSDK/latest/javadoc/com/amazonaws/services/sagemakerruntime/AmazonSageMakerRuntime.html#invokeEndpoint-com.amazonaws.services.sagemakerruntime.model.InvokeEndpointRequest-

Here is a public Github gist with examples of invocation of an endpoint from all languages including Java. https://gist.github.com/mvsusp/b9de070b310af6876de7b59beda0fb4a

I work for AWS & my opinions are my own.