yolov5_tensorrt

This is the implementation that supports the version 1.0 of yolov5s, yolov5m, yolov5l and yolov5x.
The Pytorch implementation of yolov5 version 1.0 is provided in this repos.
The latest ultralytics/yolov5 version 2.0 of https://github.com/ultralytics/yolov5 is not supported
You can see source code of yolov5s on https://github.com/wang-xinyu/tensorrtx/tree/master/yolov5
You can see Pytorch implementatuon of ultralytics/yolov5(version 1.0) on https://github.com/ultralytics/yolov5/releases/tag/v1.0

Test Environment

1. GTX1080 / Ubuntu16.04 or GTX2080Ti / Ubuntu18.04
2. cuda >= 10.0 , cudnn >= 7.0.0 , tensorrt7.0.0, nvinfer7.0.0, opencv3.3(I use opencv3.4)

How to Run

You should check the compile environment in the CMakeLists.txt for each folder(yolov5s,yolov5m,yolov5l,yolov5x).
Each folder has a readme inside, which explains how to run the models inside.
You can download pt files of ultralytics/yolov5(s,m,l,x)(version 1.0) on BaiduCloud, passwd:9cw1
Or you can download pt files of ultralytics/yolov5(s,m,l,x)(version 1.0) on https://github.com/ultralytics/yolov5/releases/tag/v1.0
Now, this repos only support yolov5 version 1.0.

Acknowledgments

Thanks to wang-xinyu for his great works and repos.