This repository contains template code for event based vision using an augmented YOLO object detector
We have used Prophesee's wonderful dataset available here: https://www.prophesee.ai/2020/01/24/prophesee-gen1-automotive-detection-dataset/
And their data visualization and extraction code available here: https://github.com/prophesee-ai/prophesee-automotive-dataset-toolbox
It is intended to run on our lab server: icsrl-exxact1.ece.gatech.edu
You will likely want to setup a virtual environment (virtualenv) for this. And you will need to append your .bashrc with:
if [ -d "/usr/local/cuda-10.1/bin/" ]; then
export PATH=/usr/local/cuda-10.1/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-10.1/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
fi
Because the default CUDA on the server is 10.2, but we need 10.1 for Tensorflow 2.2
Georgia Institute of Technology, ICSRL (http://icsrl.ece.gatech.edu/)
This project is licensed under the MIT License - see the LICENSE.md file for details