Table Of Contents
- Description
- Support Level
- Requirements
- Quickstart
- Samples
- Usage
- Known Issues
- Contributions
- Support
- License
This repository maintains sample applications designed for NVIDIA software tools integrated with Amazon Web Services (AWS).
For select demonstrations, the sample code is contained within this repository. For others, we reference and link to exceptional demonstrations available outside of this repository.
These samples are provided as community examples and are not covered by NVIDIA Enterprise Support. They are intended as reference implementations to help users get started with NVIDIA software on AWS. See Support below for how to get help.
- An active Amazon Web Services (AWS) account with permissions to create the resources used by a given sample
- Access to NVIDIA GPU instances in your target AWS region (e.g., A10G, A100, H100, H200)
- The AWS CLI installed and configured, or access to the AWS Management Console
- Additional, sample-specific prerequisites are documented in each sample's own README
- Clone this repository:
git clone https://github.com/NVIDIA/nvidia-aws-samples.git cd nvidia-aws-samples - Browse the Samples section below and pick the one that matches your use case.
- Follow the README inside that sample's directory for setup and run instructions.
Samples are organized first by use case, then by the AWS service they run on (e.g. SageMaker, Bedrock, EKS).
| Use case | Description | Status |
|---|---|---|
inference/ |
Serving and running inference | Available |
agentic/ |
Agentic workflows and tool-using applications | Available |
training/ |
Model training and fine-tuning | Planned |
data-processing/ |
Data curation and processing pipelines | Planned |
physical-ai/ |
Robotics and simulation workloads | Planned |
industry-solutions/ |
Industry-specific solutions | Planned |
Within each use case, samples are grouped by AWS service. For example, under inference/ you'll find bedrock/ and eks/, each containing self-contained samples with their own README.
Each sample directory contains its own README with detailed deployment and usage instructions. In general:
- Provision the required AWS infrastructure (cluster, compute, networking, storage) as described in the sample.
- Deploy the NVIDIA software components.
- Run the included workloads or applications.
- Clean up the resources when you are done to avoid unnecessary charges.
None at this time.
Contributions are welcome. Developers can contribute by opening a pull request and agreeing to the terms in CONTRIBUTING.md.
For questions or issues:
- Open a GitHub issue for bug reports or feature requests
- Refer to each sample's README for sample-specific guidance
To report a security vulnerability, follow the process in SECURITY.md.
See LICENSE. This project is licensed under the Apache License 2.0.