Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Prebuilt container for Ros Iron and jetson inference #286

Closed
Fibo27 opened this issue Sep 14, 2023 · 6 comments
Closed

Prebuilt container for Ros Iron and jetson inference #286

Fibo27 opened this issue Sep 14, 2023 · 6 comments

Comments

@Fibo27
Copy link

Fibo27 commented Sep 14, 2023

Hi @dusty-nv
I am visiting your repo after a while - the revamped version looks well organized - kudos for that.

Do you have a prebuilt container that i can download from the hub that has Ros iron and Jetson Inference included - I want to use it on a Orin Nano

Thanks

@dusty-nv
Copy link
Owner

@Fibo27 sorry I don't - haven't integrated that build into new jetson-containers yet. If you switch back to the legacy branch though, you could try building it with the scripts/docker_build_ros.sh --distro=iron --with-pytorch

@Fibo27
Copy link
Author

Fibo27 commented Sep 14, 2023

ok - i would presume then I have to add "iron" in line 47 of https://github.com/dusty-nv/jetson-containers/blob/bc8d0264ef25aa0d1d25a54e4658f491d2fa130f/scripts/docker_build_ros.sh - is that correct?

How about if I were to build a container using your latest script and then commit the changes using a new image name on dockerhub?

@dusty-nv
Copy link
Owner

dusty-nv commented Sep 14, 2023 via email

@jiefengsun
Copy link

jiefengsun commented Sep 17, 2023

@dusty-nv, thanks for the well-organized updates!
I would like to ask a few related questions:

  1. Is is still true: The package options are: ros_base, ros_core, and desktop - note that the ROS2 Desktop packages only build on JetPack 5.x. in the readme.md. Because I see: dustynv/ros:humble-pytorch-l4t-r32.7.1 appeared 10 days ago.. I am wondering what it is based on, (the base, core, or desktop version)?
  2. Is there a modularized way to build jetson-inference model as a package that we can select?
  3. I saw that it has jetson-inference package built in humble-pytorch-l4t-r32.7.1. Is this true for all these with pytorch tag?

I would like to have a container like: humble-desktop-pytorch-jetsoninference-l4t-r32.7.1.
With the current code, I can easily build humble-desktop-pytorch-l4t-r32.7.1, but it seems not straightforward to add the jetson-inference library.

I am not sure what is the most convenient way to do this right now since the structure of building containers changed a lot.

@dusty-nv
Copy link
Owner

  1. It's based on ros-base
  2. Not yet, but I plan on adding it. In the jetson-inference dev branch I have begun migrating it
  3. Yes, the pytorch tags actually have jetson-inference in them (in addition to pytorch)

For now, I would just revert to the jetson-containers legacy branch and you can continue building ros:*-pytorch containers the existing way. Hope that helps!

@jiefengsun
Copy link

@dusty-nv
Thanks! humble-pytorch-l4t-r32.7.1 works for me. I am able to pull it without building it locally.

I am looking forward to dev updates on building jetson inference as a package in containers. It will be a really useful function!

.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants