Skip to content

Latest commit

 

History

History
79 lines (65 loc) · 3.26 KB

JETSON_SETUP.md

File metadata and controls

79 lines (65 loc) · 3.26 KB

Jetson Nano setup

A lot of these dependencies are not pre-compiled for the Jetson Nano so expect this to take multiple hours.

1. Install Jetpack on your Jetson Nano

Jetpack is an image with Deep Learning related libs preinstalled. Follow the instructions at https://developer.nvidia.com/embedded/learn/get-started-jetson-nano-devkit

2. Setup a virtualenv

I recommend you use a virtual env to isolate this project dependencies from your system.

sudo apt install python3-pip
sudo apt install -y python3-venv
python3 -m venv ~/python-envs/rc
source ~/python-envs/rc/bin/activate

3. Install torch and torchvision

Follow the instructions at https://forums.developer.nvidia.com/t/pytorch-for-jetson-version-1-8-0-now-available/
might also need https://forums.developer.nvidia.com/t/cannot-install-pytorch/149226/3 for pytorch (from https://forums.developer.nvidia.com/t/cannot-install-pytorch/149226/3)
You may have issues installing torchvision in your virtualenv. Try this:

cd torchvision
easy_install .  --user --install-dir ~/python-envs/rc/lib/python3.6/site-packages

4. Setup TensorRT

Follow the instructions at https://docs.donkeycar.com/guide/robot_sbc/tensorrt_jetson_nano/

5. Setup the dependencies to interract with the IMU

Follow the instructions at https://docs.donkeycar.com/parts/imu/.

6. Install Jetson Stats

Follow instructions at https://github.com/rbonghi/jetson_stats

7. Link the preinstalled Opencv to your virtual env

Re-compiling cv2 takes a long time, so you can link the pre-compiled version to your virtualenv.
ln -s /usr/lib/python3.6/dist-packages/cv2/python-3.6/cv2.cpython- ite-packages/cv2.cpython-36m-aarch64-linux-gnu.so

8. Install jetracer, torch2trt and jetcam

  1. Open a terminal and call the following to install the JetCam Python package.

    cd $HOME
git clone https://github.com/NVIDIA-AI-IOT/jetcam
cd jetcam
~/python-envs/rc/bin/python setup.py install

  2. Execute the following command to install the torch2trt Python package

    cd $HOME
git clone https://github.com/NVIDIA-AI-IOT/torch2trt
cd torch2trt
~/python-envs/rc/bin/python setup.py install

  3. Execute the following in a terminal to install the JetRacer package

    cd $HOME
git clone https://github.com/NVIDIA-AI-IOT/jetracer
cd jetracer
~/python-envs/rc/bin/python setup.py install

9. If you intend to use the object detections features from this repo:

Install the yolov5 pip package pip install yolov5 You might miss libs to install/compile yolov5 dependencies:

sudo apt-get update sudo apt-get install -y build-essential libatlas-base-dev gfortran

Download yolov5s weights:

cd wandb-jetracer/src/scripts/
wget https://github.com/ultralytics/yolov5/releases/download/v5.0/yolov5s.pt