1. Install jtop
sudo apt install python3-pip
sudo -H pip3 install -U jetson-stats
reboot
After restarting, enter jtop in the terminal (ctrl+alt+t) to see various information
2. Install jetpack
sudo apt-get update
sudo apt-get dist-upgrade
sudo apt-get install nvidia-jetpack
After the installation is complete, open jtop and you can see
3. Install torch and torchvision
Install torch:
Select the version of pytorch you need in PyTorch for Jetson and download it. I installed the v1.12.0 version.
sudo apt-get install -y libopenblas-base libopenmpi-dev
pip3 install torch-1.12.0a0+2c916ef.nv22.3-cp38-cp38-linux_aarch64.whl
Open a new terminal on the desktop (ctrl+alt+t) and enter
python3
import torch
print(torch.__version__)
Check whether the installation is successful.
View version correspondence
The torch version is v1.12.0
and the corresponding torchvision version is v0.13.0
Install torchvision:
sudo apt-get install libjpeg-dev zlib1g-dev libpython3-dev libavcodec-dev libavformat-dev libswscale-dev
git clone --branch v0.13.0 https://github.com/pytorch/vision torchvision
cd torchvision
export BUILD_VERSION=0.13.0
python3 setup.py install --user
Installation takes a while.
After completion, open a new terminal on the desktop (ctrl+alt+t) and enter:
python3
import torchvision
print(torchvision.__version__)
The version of torchvision will be displayed. Continue to enter the following commands in this terminal:
python3
import torch
print(torch.cuda.is_available())
Return true
4. Configure the requirements of yolov5. Download yolov5
from github . I am using the yolov5-7 version. Open requirements and comment out lines 9, 16, and 17 with #.
python3 -m pip install --upgrade pip
cd yolov5
pip3 install -r requirements.txt
After the package is installed,
cd yolov5
python detect.py --source data/images --weight yolov5m.pt --view-img
It will be displayed in the terminal and the running results are saved in runs/detect/exp.
The environment of yolov5 has been initially set up.
Continuously updated. . .