YOLOAir algorithm library is a series of YOLO detection algorithm combination toolbox based on PyTorch, used to combine different modules to build different networks. YOLOAir is based on the YOLOv5 code framework, and synchronously adapts to the stable YOLOv5_v6.1 update and synchronizes the v6.1 deployment ecosystem.

YOLOAir has built-in model network structures such as YOLOv5, YOLOv7, YOLOX, YOLOR, Transformer, Scaled_YOLOv4, YOLOv3, YOLOv4, YOLO-Facev2, TPH-YOLOv5, YOLOv5Lite, and PicoDet, which are still being updated.

modularization: Help users to customize and quickly combine Backbone, Neck, and Head to diversify network models, help scientific research improve detection algorithms, model improvements, network arrangement and combination, and build powerful network models.

Main features

Support more YOLO series algorithm model improvements

The YOLOAir algorithm library summarizes a variety of mainstream YOLO series detection models, and a set of codes integrates multiple models:

• Built-in integrated YOLOv5 model network structure, YOLOv7 model network structure, YOLOR model network structure, YOLOX model network structure, Scaled_YOLOv4 model network structure, YOLOv4 model network structure, YOLOv3 model network structure, YOLO-FaceV2 model network structure, TPH-YOLOv5 model network structure , YOLOv5-Lite model network structure, PicoDet model network structure, etc., continuously updated…

• The above multiple detection algorithms use a unified model code framework,Integrated in the YOLOAir library, unified task form, unified application method. It is convenient for researchers to improve the algorithm model of the paper, compare the models, and realize the diversification of the network combination. Engineering algorithms are more convenient to deploy and implement, including lightweight models and higher-precision models, which can be reasonably selected according to the scene and strike a balance between accuracy and speed. At the same time, the library supports decoupling of different structures and module components to make modules componentized. By combining different module components, users can customize and build different detection models according to different data sets or different business scenarios.

use

Install

existPython>=3.7.0 Clone the repository and install the requirements.txt in your environment, includingPyTorch>=1.7.

$ git clone https://github.com/iscyy/yoloair.git  # 克隆
$ cd YOLOAir
$ pip install -r requirements.txt  # 安装

train

$ python train.py --data coco128.yaml --cfg configs/yolov5/yolov5s.yaml #默认为yolo

reasoning

detect.py Run inference on various data sources and save detection results to runs/detect Table of contents.

$ python detect.py --source 0  # 网络摄像头
                          img.jpg  # 图像
                          vid.mp4  # 视频
                          path/  # 文件夹
                          path/*.jpg  # glob

fusion

If you use different models to infer on the dataset, you can use the wbf.py file to integrate the results by weighted box fusion. You just need to set img path and txt path in wbf.py file.