Skip to content

AWS Neuron support #15172

Open
Open
@carmocca

Description

@carmocca

🚀 Feature

https://awsdocs-neuron.readthedocs-hosted.com/en/latest/index.html
https://aws.amazon.com/machine-learning/neuron/

Motivation

https://aws.amazon.com/about-aws/whats-new/2022/10/ec2-trn1-instances-high-performance-cost-effective-deep-learning-training/

Pitch

Neuron is XLA based, so it'll probably be an accelerator and strategy. Their marketing materials also advertise that it supports FSDP, and Megatron-LM. It's unclear at this point whether we would have specialized Neuron-based implementations for those.

We would also need access to hardware for testing.


If you enjoy Lightning, check out our other projects! ⚡

  • Metrics: Machine learning metrics for distributed, scalable PyTorch applications.

  • Lite: enables pure PyTorch users to scale their existing code on any kind of device while retaining full control over their own loops and optimization logic.

  • Flash: The fastest way to get a Lightning baseline! A collection of tasks for fast prototyping, baselining, fine-tuning, and solving problems with deep learning.

  • Bolts: Pretrained SOTA Deep Learning models, callbacks, and more for research and production with PyTorch Lightning and PyTorch.

  • Lightning Transformers: Flexible interface for high-performance research using SOTA Transformers leveraging PyTorch Lightning, Transformers, and Hydra.

Metadata

Metadata

Assignees

No one assigned

    Labels

    acceleratorfabriclightning.fabric.FabricfeatureIs an improvement or enhancementplGeneric label for PyTorch Lightning packagestrategy

    Type

    No type

    Projects

    No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions