Official Inquiry: Is 'jetson-containers' the Recommended Production Workflow for JetPack 6.x/7.x on Ubuntu?

Robotics & Edge Computing Jetson Systems Jetson Thor
1

Dear NVIDIA Jetson Team,

I am writing to seek an official recommendation regarding the use of jetson-containers as the primary development and deployment framework for our commercial products based on Ubuntu L4T.

As we transition our AI stack to JetPack 6.x (Ubuntu 22.04) and prepare for the Blackwell-based Jetson Thor (JetPack 7.x), we are evaluating whether to standardize our pipeline on the jetson-containers build system. While it is highly efficient for integrating LLM/VLM stacks, we need to clarify its status as a “Standard NVIDIA Product” versus a “Developer Reference Tool.”

Specifically, I would like to confirm:
Production Recommendation: Does NVIDIA officially recommend jetson-containers for building and maintaining production-grade container images on Ubuntu? Or is the official guidance still limited to using raw images from NGC (NVIDIA GPU Cloud)?
Support for L4T 6.x/7.x: Given the major kernel and OS shifts in JetPack 6 and 7, will jetson-containers receive synchronized updates and LTS support alongside the official L4T releases?
Blackwell Acceleration (Thor): For Jetson Thor, will the optimization scripts within jetson-containers be validated by the NVIDIA product team to ensure full utilization of Blackwell’s hardware acceleration features?
Security & Compliance: Since production environments require strict security scanning, are the images generated by jetson-containers audited by NVIDIA for vulnerabilities, similar to the verified images on NGC?

Thank you for your guidance.

3

Hi,

It’s recommended to use our NGC container.
For example:

vLLM: vLLM | NVIDIA NGC
SGLang: SGLang | NVIDIA NGC

Thanks.

4

Dear Aasta,

I am currently evaluating the development workflow for our AI applications and would appreciate your guidance on the following:

  1. Development & Encapsulation: Are there any official documents or best practices regarding the development and encapsulation of AI applications using Docker?

  2. NGC Image Customization: Is it possible to perform secondary development on images from the NGC Catalog? Specifically, can these be used as base images for further customization and re-packaging?

Looking forward to your professional advice.

5

Hi,

1. You can find some examples and tutorials below:

2. Yes.

Thanks.

7

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.