YOLOSHOW - YOLOv5 / YOLOv7 / YOLOv8 / YOLOv9 / YOLOv10 / RTDETR GUI based on Pyside6

Introduction

YOLOSHOW is a graphical user interface (GUI) application embed withYOLOv5 YOLOv7 YOLOv8 YOLOv9 YOLOv10 RT-DETR algorithm.

English | 简体中文

Demo Video

YOLOSHOW v1.x : YOLOSHOW-YOLOv9/YOLOv8/YOLOv7/YOLOv5/RTDETR GUI

YOLOSHOW v2.x : YOLOSHOWv2.0-YOLOv9/YOLOv8/YOLOv7/YOLOv5/RTDETR GUI

Todo List

Functions

1. Support Image / Video / Webcam / Folder (Batch) / IPCam Object Detection

Choose Image / Video / Webcam / Folder (Batch) / IPCam in the menu bar on the left to detect objects.

2. Change Models / Hyper Parameters dynamically

When the program is running to detect targets, you can change models / hyper Parameters

Support changing model in YOLOv5 / YOLOv7 / YOLOv8 / YOLOv9 / RTDETR / YOLOv5-seg / YOLOv8-seg / YOLOv10 dynamically
Support changing IOU / Confidence / Delay time / line thickness dynamically

3. Loading Model Automatically

Our program will automatically detect pt files including YOLOv5 Models / YOLOv7 Models / YOLOv8 Models / YOLOv9 Models / YOLOv10 Models that were previously added to the ptfiles folder.

If you need add the new pt file, please click Import Model button in Settings box to select your pt file. Then our program will put it into ptfiles folder.

Notice :

All pt files are named including yolov5 / yolov7 / yolov8 / yolov9 / yolov10 / rtdetr . (e.g. yolov8-test.pt)
If it is a pt file of segmentation mode, please name it including yolov5n-seg / yolov8s-seg . (e.g. yolov8n-seg-test.pt)
If it is a pt file of pose estimation mode, please name it including yolov8n-pose . (e.g. yolov8n-pose-test.pt)
If it is a pt file of oriented bounding box mode, please name it including yolov8n-obb . (e.g. yolov8n-obb-test.pt)

4. Loading Configures

After startup, the program will automatically loading the last configure parameters.
After closedown, the program will save the changed configure parameters.

5. Save Results

If you need Save results, please click Save Mode before detection. Then you can save your detection results in selected path.

6. Support Object Detection, Instance Segmentation and Pose Estimation

From YOLOSHOW v3.0，our work supports both Object Detection , Instance Segmentation, Pose Estimation and Oriented Bounding Box. Meanwhile, it also supports task switching between different versions，such as switching from YOLOv5 Object Detection task to YOLOv8 Instance Segmentation task.

7. Support Model Comparison among Object Detection, Instance Segmentation, Pose Estimation and Oriented Bounding Box

From YOLOSHOW v3.0，our work supports compare model performance among Object Detection, Instance Segmentation, Pose Estimation and Oriented Bounding Box.

Preparation

Experimental environment

OS : Windows 11 
CPU : Intel(R) Core(TM) i7-10750H CPU @2.60GHz 2.59 GHz
GPU : NVIDIA GeForce GTX 1660Ti 6GB

1. Create virtual environment

create a virtual environment equipped with python version 3.9, then activate environment.

conda create -n yoloshow python=3.9
conda activate yoloshow

2. Install Pytorch frame

Windows: pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
Linux: pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

Change other pytorch version in

3. Install dependency package

Switch the path to the location of the program

cd {the location of the program}

Install dependency package of program

pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple

4. Add Font

Windows User

Copy all font files *.ttf in fonts folder into C:\Windows\Fonts

Linux User

mkdir -p ~/.local/share/fonts
sudo cp fonts/Shojumaru-Regular.ttf ~/.local/share/fonts/
sudo fc-cache -fv

MacOS User

The MacBook is so expensive that I cannot afford it, please install .ttf by yourself. 😂

5. Run Program

python main.py

waketzheng/YOLOSHOW