Skip to content

Latest commit

 

History

History
89 lines (59 loc) · 2.64 KB

README-en.md

File metadata and controls

89 lines (59 loc) · 2.64 KB
Raw

TensorRT C++ api deploy YOLOv8 + ByteTrack

Introduction

  • Based on TensorRT-v8 , deploy YOLOv8 + ByteTrack ;

  • Support Jetson series, also Linux x86_64;

Main work I have done:

  1. Refer to tensorrtx ，model: .pth -> .engine,extract the inference part of the code, encapsulated into C++ classes, easy to call other projects ;
  2. Preprocessing replaced with my own CUDA programming preprocessing;
  3. Post-processing removed CUDA programming because it was not significantly faster in tests compared to CPU post-processing ;
  4. The post-processed NMS greatly reduces conf_thres hyperparameters due to the principle of ByteTrack tracking, which is very important ;
  5. YOLOv8 inference compiles to a dynamic link library to decouple projects;
  6. Reference official ByteTrack TensorRT deploy , modify its interface to the YOLO detector;
  7. ByteTrack also compiles to a dynamic link library, further decoupling projects;
  8. Add category filtering function, you can set the category you want to track in main.cpp line 8 .

Effect

Environment

  1. Base requirements：
  • TensorRT 8.0+
  • OpenCV 3.4.0+
  1. My running environment on Jetson Nano is as follows:
  • The burned system image is Jetpack 4.6.1，original environment is as follows：
CUDA cuDNN TensorRT OpenCV
10.2 8.2 8.2.1 4.1.1
  • Install Eigen
apt install libeigen3-dev

Model conversion

Get the serialized file of TensorRT, suffix.engine

  • First get the wts format model file, link: yolov8s.wts , code：gsqm
  • Then follow these steps:
cd {TensorRT-YOLOv8-ByteTrack}/tensorrtx-yolov8/
mkdir build
cd build
cp {path/to/yolov8s.wts} .
cmake ..
make
./yolov8 -s yolov8s.wts yolov8s.engine s

cd ../../
mkdir yolo/engine
cp tensorrtx-yolov8/build/yolov8s.engine yolo/engine

Run tracking

  • Follow these steps
mkdir build
cd build
cmake ..
make
./main ../videos/demo.mp4  # The path to your own video

Then the result.mp4 will be in the build directory, is to track the effect of the video file

If you want the tracked video to play in real time, you can uncomment line 94 of main.cpp.

Reference