Bring Your Own Neuron Container to SageMaker Hosting (Inf2 or Trn1) — Archived

Warning

This document is archived. The Inf2 and Trn1 SageMaker BYOC developer flow is no longer maintained as a standalone page. It is provided here for reference only.

The original page described how to compile a model on an EC2 instance or a SageMaker Notebook, then deploy it to SageMaker Hosting on ml.inf2 or ml.trn1 using a custom container.

For current guidance, see:

• Compile with Framework API and Deploy on EC2 Inf2 for the EC2 Inf2 deployment flow.

• The AWS guide to Adapting Your Own Inference Container for the SageMaker Hosting BYOC pattern.

• Tutorial How to Build and Run a Neuron Container for building a Neuron container image.

• The Compiling and Deploying HuggingFace Pretrained BERT on Inf2 on Amazon SageMaker sample for an end-to-end Inf2 BYOC example.