T1M-CHEN/Pytorch_Mask

人脸口罩佩戴检测(识别)的AI Pipeline优化与部署

Mask/No Mask Wearing Detection and Optimization with YOLOv5

---

Introduction

Learning to build a project, including:

• Using YOLOv5 to detect mask wearing.
• Using tensorrtx to optimize the model.
• Using flask to display the results.

Version

YOLOv5-5.0

TensorRT 7.2.2.3

Training

• Release resources in

Setup & Environments

Intel E5 *5 Cores +GTX 1050Ti / TITAN X

Ubuntu 18.04 + CUDA 11.0 + CUDNN 8

Python 3.8 + Pytorch 1.7.1

Validation

YOLOv5 benchmark:

python val.py --data ./voc-mask.yaml --weights weights/yolov5s_best.pt --batch-size 16

Flask Deployment:

• Install Flask
• Run Flask：sh run.sh
• Access 127.0.0.1:8888

References

YOLOv5

tensorrtx