We strongly recommend you to run the following commands from the system shell or terminal.
DO NOT RUN these commands from any IDE terminal. There are some strange issues especially from PyCharm Terminal
1.1 Clone the repository from github
git clone git@github.com:YiweiC1W/pedestrians_detection.git
git clone https://github.com/YiweiC1W/pedestrians_detection.git
1.2 cd to the directory
cd pedestrians_detection
Unzip the folder
cd into the project root folder (where the README.md is)
2.1 Create virtual environment
conda create --name 9517gp python==3.8.13
2.2 Activate the virtual environment
conda activate 9517gp
3.Run this All-in-One script to install dependencies, YOLOX, Download datasets and pre-trained weights (choose one according to your system)
sh install_linux.sh # For Linux (Ubuntu 20.04 is recommended) (maybe works on macOS too)
install_win64.bat # For Windows (not tested)
If you encounter any installation issues, please contact me at yiwei.chen2@student.unsw.edu.au
please edit IMAGE_FOLDER_PATH in task1.py
or you can use args '--path' to choose your image folder
'--task' to choose your task eg: '--task 1' or '--task 2'
'--conf' to choose your confidence threshold
'--device' to choose your device 'cpu' or 'gpu'
'--video' bool(true or false), if you want to save video
'--picture' bool(true or false), if you want to save picture
python main.py --task 1 --device cpu