ubuntu 下安装 TensorRT 7.0 以及验证

安装环境

Ubuntu 18.04
TensorRT-7.0.0.11
cuda 10.0
cudnn 7.6.5

• 安装 TensorRT 7.0 版本

  Tar File Installation
  This section contains instructions for installing TensorRT from a tar file.
  About this task

  Note: Before issuing the following commands, you’ll need to replace 7.x.x.x with your specific TensorRT version. The following commands are examples.

  Procedure

  1. Install the following dependencies, if not already present:
     CUDA 10.0
     cuDNN 7.6.5
     Python 3 (Optional)

  2. Download the TensorRT tar file that matches the Linux distribution you are using.

  3. Choose where you want to install TensorRT. This tar file will install everything into a subdirectory called TensorRT-7.x.x.x.

  4. Unpack the tar file.

     version="7.x.x.x"  
     os=""  
     arch=$(uname -m)      
     cuda="cuda-x.x"   
     cudnn="cudnn8.x"   
     tar xzvf TensorRT-${version}.${os}.${arch}-gnu.${cuda}.${cudnn}.tar.gz  
     

     Where:

     • 7.x.x.x is your TensorRT version
     • is:
       Ubuntu-16.04
       Ubuntu-18.04
       CentOS-7.6
     • cuda-x.x is CUDA version 10.2 or 11.0.
     • cudnn8.x is cuDNN version 8.0.
       This directory will have sub-directories like lib, include, data, etc…

     ls TensorRT-${version}
     bin  data  doc  graphsurgeon  include  lib  python  samples  targets  TensorRT-Release-Notes.pdf  uff
     

  5. Add the absolute path to the TensorRTlib directory to the environment variable LD_LIBRARY_PATH:

     export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:
     

  6. Install the Python TensorRT wheel file.

     cd TensorRT-${version}/python
     

     If using Python 2.7:

     sudo pip2 install tensorrt-*-cp27-none-linux_x86_64.whl
     

     If using Python 3.x:

     sudo pip3 install tensorrt-*-cp3x-none-linux_x86_64.whl
     

  7. Install the Python UFF wheel file. This is only required if you plan to use TensorRT with TensorFlow.

     cd TensorRT-${version}/uff
     

     If using Python 2.7:

     sudo pip2 install uff-0.6.9-py2.py3-none-any.whl
     

     If using Python 3.x:

     sudo pip3 install uff-0.6.9-py2.py3-none-any.whl
     

     In either case, check the installation with:

     which convert-to-uff
     

  8. Install the Python graphsurgeon wheel file.

     cd TensorRT-${version}/graphsurgeon
     

     If using Python 2.7:

     sudo pip2 install graphsurgeon-0.4.5-py2.py3-none-any.whl
     

     If using Python 3.x:

     sudo pip3 install graphsurgeon-0.4.5-py2.py3-none-any.whl
     

  9. Verify the installation:

     a. Ensure that the installed files are located in the correct directories. For example, run the tree -d command to check whether all supported installed files are in place in the lib, include, data, etc… directories.
     b. Build and run one of the shipped samples, for example, sampleMNIST in the installed directory. You should be able to compile and execute the sample without additional settings. For more information, see the “Hello World” For TensorRT (sampleMNIST).
     c. The Python samples are in the samples/python directory.

• 编译TensorRT的例子 (sampleMNIST) 出错：

  ~/Downloads/TensorRT/TensorRT-7.0.0.11/samples/sampleMNIST$ make
  ../Makefile.config:7: CUDA_INSTALL_DIR variable is not specified, using /usr/local/cuda by default, use CUDA_INSTALL_DIR=<cuda_directory> to change.
  ../Makefile.config:10: CUDNN_INSTALL_DIR variable is not specified, using $CUDA_INSTALL_DIR by default, use CUDNN_INSTALL_DIR=<cudnn_directory> to change.
  make: Nothing to be done for 'all'.
  

  解决办法：

  vim ~/.bashrc
  
  # tensorrt cuda and cudnn
  export CUDA_INSTALL_DIR=/usr/local/cuda
  export CUDNN_INSTALL_DIR=/usr/local/cuda
  

  ---

  :~/Downloads/TensorRT/TensorRT-7.0.0.11/samples/sampleMNIST$ make -j8
  make: Nothing to be done for 'all'.
  

  解决办法：

  make clean
  

  ---

  ~/Downloads/TensorRT/TensorRT-7.0.0.11/bin$ ./sample_mnist
  ./sample_mnist: error while loading shared libraries: libnvinfer.so.7: cannot open shared object file: No such file or directory
  
  

  解决办法：

  vim ~/.bashrc
  export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/xxx/Downloads/TensorRT/TensorRT-7.0.0.11/lib
  

  ---

  ~/Downloads/TensorRT/TensorRT-7.0.0.11/bin$ ./sample_mnist
  &&&& RUNNING TensorRT.sample_mnist # ./sample_mnist
  [07/14/2020-11:43:08] [I] Building and running a GPU inference engine for MNIST
  [07/14/2020-11:43:10] [I] [TRT] Detected 1 inputs and 1 output network tensors.
  [07/14/2020-11:43:10] [W] [TRT] Current optimization profile is: 0. Please ensure there are no enqueued operations pending in this context prior to switching profiles
  Could not find 6.pgm in data directories:
      data/mnist/
      data/samples/mnist/
  &&&& FAILED
  

  解决办法：

  cd /TensorRT-7.0.0.11/data/mnist
  python download_pgms.py 
  

参考：
install and configure tensorrt 4 on ubuntu 16.04
【TensorRT】tensorRT 7.0的安装配置