This document is relevant for: Trn2, Trn3

Legacy training frameworks

Find the existing reference documentation for the PyTorch/XLA training path and its distributed training libraries:

• NxD Training and PyTorch/XLA training: Library overview, setup, tutorials, developer guides, and API references for NxD Training and the XLA-based torch-neuronx training flow.

• NxD Core for training: Distributed training primitives, setup, developer guides, and API references.

These links point to the existing documentation; see each destination for its support status. For current training options, see Train models on Neuron.

For inference documentation, see Legacy inference frameworks.

This document is relevant for: Trn2, Trn3