This document is relevant for: Inf1

Warning

This document is archived. MXNet is no longer officially supported by the AWS Neuron SDK. It is provided for reference only. For current framework support, see ML framework support on AWS Neuron SDK.

Compile with Sagemaker Neo and Deploy on Sagemaker Hosting (inf1)#

Description#

Neuron developer flow on SageMaker Neo

You can use SageMaker Neo to compile models for deployment on SageMaker Hosting using ml.inf1 instances. In this developer flow, you provision a Sagemaker Notebook instance to train, compile and deploy your model using the SageMaker Python SDK. Follow the steps bellow to setup your environment.

Setup Environment#

  1. Create an Amazon SageMaker Notebook Instance:

    Follow the instructions in Get Started with Notebook Instances

    The Notebook instance created provides the required Python SDK for training, compiling and deploying models with Amazon SageMaker.

  2. Compile a model using the Amazon SageMaker SDK:

    Refer to Supported Instances Types and Frameworks for information on the framework versions currently supported by Amazon SageMaker Neo on AWS Inferentia.

    More information about compiling and deploying models with Amazon SageMaker Neo can be found on Use Neo to Compile a Model

This document is relevant for: Inf1