TIPS

This repo is based on YOLOv5 v6.1

Download the models of YOLOv5 v6.1 from here，and put the them to the pt folder. When the GUI runs, the existing models will be automatically detected.

Other versions: v5.0, ...

If you're not interested in code, you can use the packaged GUI. Download link: Google Drive or Baidu Netdisk-pwd=6666.

Download the zip file and unzip it to the destination, find and run main.exe

Updated Date：2023/2/1

Demo Video：

https://www.bilibili.com/video/BV1sQ4y1C7Vk?spm_id_from=333.999.0.0

Quick Start

conda create -n yolov5_pyqt5 python=3.8
conda activate yolov5_pyqt5
pip install -r requirements.txt
python main.py

About Packaging

• install pyinstaller

pip install pyinstaller==5.7.0

• package the GUI

pyinstaller -D -w --add-data="./utils/*;./utils" --add-data="./config/*;./config" --add-data="./icon/*;./icon" --add-data="./pt/*;./pt" --add-data="./imgs/*;./imgs" main.py

• if no errors occur, the packaged application is in dist/main

Function

1. support image/video/webcam/rtsp as input
2. change model
3. change IoU
4. change confidence
5. set latency
6. paly/pause/stop
7. result statistics
8. save detected image/video automatically

You can find ui files in main_win and dialog

Further Improvements

• YOLOv8-PySide6-GUI by Jai-wei

• YOLOv5_v7.0 + deepsort by CUMTduchen