THE PROCESS
ssh nano@192.168.86.32
cd Development/vexai8029_org # this directory is used as a docker volume when the docker container is run
git pull origin main # get the most recent updates from this repo
sudo docker run --gpus all -it --privileged -v /dev:/dev -v /home/nano/Development/vexai8029_org/:/usr/shared-dev yolov5:1.0
#the model should automatically start after launching the docker container
Vex-AI-2021-Code
Setting up ssh on the vex administered jetson: sudo nano /etc/NetworkManager/NetworkManager.conf then change managed=false to managed=true sudo service network-manager restart then connect to wifi, safe ip, and ssh.
Clone repository
Put copy runs folder from google drive into the yolov5 folder
run detectVexai.py from WITHIN the yolov5 folder
you may have to tweak the camera parameter
(this model was trained on a small dataset for a very limited time period, so take all results with a grain of salt)
REQUIRED DEPENDANCIES FOR PYTORCH / TORCHVISION
#RUN apt install nvidia-cuda-toolkit
#sudo apt-get install libavformat-dev
#sudo apt-get install libavcodec-dev
#sudo apt-get install libswscale-dev
#sudo apt-get install libjpeg-dev
#pip3 install pillow
#pip3 install tqdm
#pip3 install matplotlib
#pip3 install pandas
#pip3 install seaborn
HOW TO DOWNLOAD FILES FROM GOOGLE DRIVE VIA COMMAND LINE
sudo apt-get install unzip
wget --no-check-certificate 'https://docs.google.com/uc?export=download&id={YOURIDHERE}' -r -A 'uc*' -e robots=off -nd
{YOURIDHERE} = digits in url after /d/
ls -l to find the largest of the three created files
mv uc?export=download&confirm=w3R-&id=19aByVTPC1xEMTIuDg7-h__tl2Iy-Bo5U~ runs. zip change the name to the appropriate name and filetype
unzip runs.zip
done :)
to run a dockercontainer on startup, use the docker --restart=always flag
another option is to use systemd like this: https://askubuntu.com/questions/620930/how-do-i-autostart-docker-container-at-system-reboot
more info (https://docs.docker.com/config/containers/start-containers-automatically/)