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

Jetson inference #122

Open
vijayreddysamula opened this issue May 20, 2023 · 5 comments
Open

Jetson inference #122

vijayreddysamula opened this issue May 20, 2023 · 5 comments

Comments

@vijayreddysamula
Copy link

Hi, Dusty,
I am working Jetson Orin AGX developer kit.
I encountered a small problem after building the ros image. The container doesn't have jetson inference and jetson utils. But I tried cloning the jetson inference and it always shows webrtc error. I don't want to build a jetson inference again, so I wrote the docker file with the jetson inference image as the base and installed ros2. But there I can see the ros2 topic list but not able to subscribe.

@dusty-nv
Copy link
Owner

Hi @vijayreddysamula, which version of jetpack-l4t are you running? (you can check this with cat /etc/nv_tegra_release)

You can build the ros:pytorch container (which has ros_deep_learning installed and uses jetson-inference as base) with jetson-containers scripts/docker_build_ros.sh --distro=humble --with-pytorch

@vijayreddysamula
Copy link
Author

Hi Dusty,
I am using L4t r34.1.1. I tried jetson-containers with the above script using foxy, but i got following error pull access denied for jetpack, repository does not exist or may require 'docker login': denied: requested access to the resource is denied

@dusty-nv
Copy link
Owner

Hi @vijayreddysamula, IIRC the l4t-jetpack base container for L4T R34.1.1 wasn't released - L4T R34.1.1 was a Developer Preview release and you should upgrade it (i.e. to the latest - currently JetPack 5.1.1 / L4T R35.3.1)

@jiefengsun
Copy link

Hi @vijayreddysamula, which version of jetpack-l4t are you running? (you can check this with cat /etc/nv_tegra_release)

You can build the ros:pytorch container (which has ros_deep_learning installed and uses jetson-inference as base) with jetson-containers scripts/docker_build_ros.sh --distro=humble --with-pytorch

@dusty-nv This means we can directly do inference in the same node by calling imagenet API from jetson-interface. Is this a solution to avoid nodelet?

@dusty-nv
Copy link
Owner

@jiefengsun in ROS2 it doesn't use nodelets, but yes. Isaac ROS also has zero-copy transport between nodes (NITROS) https://github.com/NVIDIA-ISAAC-ROS/isaac_ros_nitros

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