directory
I. Environment version
1.1 an overview of the
Ubuntu 1.2 version
1.3 Python, PyTorch, and Tensorrt
Version 1.4 cudnn
1.5 Driver Version
Opencv version 1.6
TensorRT and OpencV environment installation
Opencv installed 2.1
2.2 TensorRT installation
Yolov5 to tensorrt
3.1 Downloading the YoloV5 Project
3.2 Downloading the Tensorrtx project
3.3 Preparing for Compilation
3.4 Obtaining the WTS model
3.5 Compiling Commands
3.6 test
I. Environment version
1.1 an overview of the
System environment: Ubuntu18.04 Python =3.8.5 Torch =1.10.0 TensorRT=7.2.3.4 OpencV =4.5.1
Cuda =11.2 CUDNN =8.2.2 Driver Version: 460.32.03 Graphics card: RTX 3090
Ubuntu 1.2 version
Run on a terminal: lsb_release -a
1.3 Python, PyTorch, and Tensorrt
Run on a terminal: lsb_release -a
Version 1.4 cudnn
Run on terminal: DPKG -l libcudnn8
1.5 Driver Version
Run in terminal: watch -n 1 Nvidia-SMI
Opencv version 1.6
Running on the terminal: pkg-config –modversion opencV
TensorRT and OpencV environment installation
Opencv installed 2.1
- Please refer to home-OpencV
- Ippicv package: github.com/opencv/open…
- Cmake -d CMAKE_INSTALL_PREFIX=/usr/local -d CMAKE_BUILD_TYPE=Release -d BUILD_TIFF=ON
If pyTorch is installed in conda, libtiff will be installed automatically. If PyTorch is installed in Conda, libtiff will be installed automatically.
/ usr/local/lib/libopencv_imgcodecs. So the 4.5.1: Undefined reference to 'TIFFReadRGBAStrip@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFReadDirectory@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFWriteEncodedStrip@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFIsTiled@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFWriteScanline@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFGetField@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFScanlineSize@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFWriteDirectory@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFReadEncodedTile@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFReadRGBATile@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFClose@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFClientOpen@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFRGBAImageOK@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFOpen@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFReadEncodedStrip@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFSetField@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to 'TIFFSetWarningHandler@LIBTIFF_4.0' /usr/local/lib/libopencv_imgcodecs.so.4.5.1: Undefined reference to ` TIFFSetErrorHandler@LIBTIFF_4.0Copy the code
Solutions:
- When compiling opencV source code, ensure that conda environment does not have libtiff package name
- BUILD_TIFF is enabled during cmake compilation
Reference links:
- Libtiff_4. 0 link Errors – OpenCV Q&A Forum
- Stackoverflow.com/questions/2…
- OpenCV libtiff_4. 0 link errors
- Install OpenCV + OpenCV_Contrib in Ubuntu 18.04
- Blog.51cto.com/u_14587161/…
2.2 TensorRT installation
- Can refer to the network: developer.nvidia.com/zh-cn/tenso… , there are tar package and DEB two installation parties, this paper chooses tar package for decompression installation.
- Decompress the tar package and install it: tar -xvf tensorrt-7.2.3.4.ubuntu-18.04.x86_64 -gnu. Cuda-11.1.cudnn8.1.tar. gz
- Install the packages in python, uFF, and GraphSurgeon respectively
- Vi ~ /. Bashrc add export LD_LIBRARY_PATH = / usr/local/TensorRT – 7.2.3.4 / lib: / usr/local/cuda/bin: LDLIBRARYPATH exportINCLUDEPATH = / usr/local/TensorRT – 7.2.3.4 / include: / usr/local/cuda/lib64: LD_LIBRARY_PATH \ export INCLUDE_PATH = / usr/local/TensorRT – 7.2.3.4 / include: / usr/local/cuda/lib64: LDLIBRARYPATH ExportINCLUDEPATH = / usr/local/TensorRT – 7.2.3.4 / include: / usr/local/cuda/lib64: INCLUDE_PATH
- source ~/.bashrc
Yolov5 to tensorrt
3.1 Downloading the YoloV5 Project
Github.com/ultralytics…
3.2 Downloading the Tensorrtx project
Github.com/wang-xinyu/…
3.3 Preparing for Compilation
In the Tensorrx project, go to the yolov5 directory and modify cmakelists.txt:
Modify your tcUDA and TensorRT paths: # cuda include_directories(/usr/local/cuda/include) link_directories(/usr/local/cuda/lib64) # tensorrt Include_directories (/ usr/local/TensorRT - 7.2.3.4 / include) link_directories (/ usr/local/TensorRT - 7.2.3.4 / lib)Copy the code
3.4 Obtaining the WTS model
Run on a terminal: ****python gen_wts.py
3.5 Compiling Commands
Run on terminal: mkdir build && CD build:
Run on terminal: cmake..
Run on terminal: make-j128
Run on terminal:./ yolov5-s // serialize Model to Plan file i.e. ‘yolov5s.engine’ serialization model
3.6 test
Running on terminal:./ yolov5-d.. /data/images // deserialize plan file and run inference, the images in samples will be processed. The test case
Or run CUDA_VISIBLE_DEVICES=6 Python yolov5_trt.py on a terminal
