# Local AI Inference Recipes
This repository contains recommended recipes for running models on NVIDIA local hardware. Each recipe contains the exact model checkpoint, inference backend and recommended settings.
⚠️ Note this is a work in progress and is actively being developed.
Files are Apache-2.0 licensed (see `LICENSE`) unless otherwise specified
in the file — for example, the `.yaml` and `.html` files are NVIDIA
Proprietary.
## Run a recipe
1. Select your model and exact hardware configuration.
2. Review the prerequisites, limitations, and safety guidance.
3. Set any required environment variables, such as a scoped Hugging Face token.
4. Copy and run the setup command, when required.
5. Copy and run the launch command.
6. Use the provided client example to test the endpoint.
## Recommended configurations
`recommended_catalog.json` lists a curated set of standout recipes, grouped by backend and operating system.
## Feedback
This repository does not accept external contributions (no pull requests).
If a recipe appears incorrect or outdated, or you would like to request support for another model or hardware configuration, please open an issue.
## Notice and disclaimer
NOTICE AND DISCLAIMER: This interface presents the option of accessing AI models. Those materials are governed solely by separate terms, conditions and licenses. You are solely responsible for finding, reviewing and complying with all applicable terms, conditions, and licenses, and for verifying the security, integrity and suitability of any retrieved materials for your specific use case. They are presented "AS IS", without warranty of any kind. The author makes no representations or warranties regarding any retrieved materials, and assumes no liability for any losses, damages, liabilities or legal consequences from your use or inability to use any retrieved materials. Use the retrieved materials at your own risk.