Neuron application notes

Warning

This content is archived and no longer actively maintained. These application notes are provided for reference only and may describe unsupported features, deprecated APIs, or older SDK releases. For current documentation, see the AWS Neuron SDK documentation.

Application notes provide specific documentation for support announcements, migration guides, performance optimization techniques, and framework-specific implementations for AWS Neuron SDK components.

Framework integration

PyTorch Neuron (Inf1)

R-CNN implementation and optimization techniques for PyTorch on Inf1

PyTorch NeuronX Graph Partitioner

Advanced graph partitioning strategies for distributed training and inference

Data Parallel Inference on Torch NeuronX

Guide to using torch.neuronx.DataParallel for scalable inference on Inf1

Data Parallel Inference on Torch Neuron

Guide to using torch.neuron.DataParallel for scalable inference on Inf1

Migrate from XLA_USE_BF16/XLA_DOWNCAST_BF16

Guide to migrating from deprecated XLA environment variables to recommended PyTorch mixed-precision options on NeuronX

PyTorch 2.9 Support

New features and migration guide for PyTorch 2.9 on Neuron