lakms123456/orin repository overview

⁠Image Info

• 4.0-devel-jp5.1.2 (stereolabs/zed:4.0-devel-jetson-jp5.1.2⁠)

  • 4.0-devel-jp5.1.2-FenceAvoidance
    • zed-sdk
  • 4.0-devel-jp5.1.2-ROS
    • Jetpack 5.1.2 , python 3.8 , OpenCV 4.5.4 , OpenCV-python 4.2.0
      • Noetic, Zed wrapper

• R35.2.1-yolov5-MAVROS-FastPlanner-stable (nvcr.io/nvidia/l4t-pytorch:r35.2.1-pth2.0-py3⁠)

  • Based on R35.2.1-yolov5-MAVROS image
  • R35.2.1-yolov5-MAVROS-FenceAvoidance
    • Fence Avoidance, YOLOv5-custom
  • R35.2.1-yolov5-MAVROS-FastPlanner-stable
    • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.2.0 , OpenCV-python 4.2.0 , torch 2.0
      • Noetic , MAVORS , Realsense SDK (2.54.2) & Realsense ROS
      • cv_bridge (opencv4)⁠

• R35.2.1-yolov5-MAVROS (nvcr.io/nvidia/l4t-pytorch:r35.2.1-pth2.0-py3⁠)

  • Based on R35.2.1-yolov5-ROS-stable image
  • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.5.0 , OpenCV-python 4.2.0 , torch 2.0
    • Noetic , MAVORS , cv_bridge (1.16.2) , Realsense SDK (2.50.0) & ROS (2.3.2)
    • YOLOv5⁠, YOLOv5-ROS⁠

• R35.2.1-yolov5-ROS-stable (nvcr.io/nvidia/l4t-pytorch:r35.2.1-pth2.0-py3⁠)

  • Based on R35.2.1-yolov7-ROS-stable image
  • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.5.0 , OpenCV-python 4.2.0 , torch 2.0
    • Noetic , YOLOv5⁠, YOLOv5-ROS⁠

• R35.2.1-yolov7-ROS-stable (nvcr.io/nvidia/l4t-pytorch:r35.2.1-pth2.0-py3⁠)

  • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.5.0 , OpenCV-python 4.2.0 , torch 2.0
    • Noetic , YOLOv7⁠ , YOLOv7-ROS⁠

• R35.2.1-yolov7 (dustynv/jetson-inference:r35.2.1)

  • R35.2.1-yolov7-ROS
    • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.5.0 , torch 2.0
    • YOLOv7⁠ , YOLOv7-ROS⁠
    • cv_bridge (1.16.2) , Realsense SDK (2.50.0) & ROS (2.3.2)
  • R35.2.1-yolov7-base
    • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.5.0 , torch 2.0
    • YOLOv7⁠

• yolov5_noetic_ros (dustynv/ros:noetic-pytorch-l4t-r35.2.1)

  • Unstable (OpenCV)
  • Jetpack 5.1 , CUDA 11.4 , python 3.8 , OpenCV 4.7.0.72 , torch 1.13.0
  • YOLOv5⁠, YOLOv5-ROS⁠

⁠Create Docker Image (dustynv/jetson-inference:r35.2.1)

dusty-nx/jetson-inference : here⁠

📌 dustynv/jetson-inference:r35.2.1
Jetpack 5.1 (L4T R35.2.1)
CUDA 11.4, python 3.8, OpenCV 4.5.0
torch       2.0.0a0+ec3941ad.nv23.2
torchaudio  0.13.1+b90d798
torchvision 0.14.1a0+5e8e2f1  

⁠1. Modify run file

git clone --recursive --depth=1 https://github.com/dusty-nv/jetson-inference

jetson-inference/docker/run.sh

# run the container
if [ $ARCH = "aarch64" ]; then

	# /proc or /sys files aren't mountable into docker
	cat /proc/device-tree/model > /tmp/nv_jetson_model

	sudo docker run --runtime nvidia -it \
		--network host \
		-v /tmp/argus_socket:/tmp/argus_socket \
		-v /etc/enctune.conf:/etc/enctune.conf \
		-v /etc/nv_tegra_release:/etc/nv_tegra_release \
		-v /tmp/nv_jetson_model:/tmp/nv_jetson_model \
		-v /dev:/dev \
		--privileged \
		-v ~/docker_share:/root/share:rw \
		-h noetic \
		--name noetic \
		$DISPLAY_DEVICE $V4L2_DEVICES \
		$DATA_VOLUME $USER_VOLUME $DEV_VOLUME \
		$CONTAINER_IMAGE $USER_COMMAND

⁠2. Create image

cd jetson-inference
docker/run.sh --ros=noetic