TensorRT for Multi-model

notice

The original code is from wang-xinyu/tensorrtx
This code utilizes TensorRT to simultaneously accelerate both YOLOP and YOLOv11 models, deploying them on NVIDIA AGX Orin devices.

Test Environment

1. TensorRT 8.5.2
2. Nvidia AGX Orin(Jetson L4T R35.4.1)
3. CUDA 11.4
4. cuDNN 8.6.0
5. OpenCV 4.5.4

how to use

1. build yolop-yolo11

   cd yolop-yolo11
   mkdir build && cd build
   cmake ..
   make

3. run

   ./multi-model

Acknowledgments

This project incorporates code from the following third-party source:

• Source: WangXinyu/original-repo
• License: MIT License
• Copyright: © 2019-2020 Wang Xinyu
• Usage: [Implement multi-task detection of lane lines and targets]