Installation, download and sharing of LabVIEW graphical TensorRT toolkit

Preface

Today we will take a look at how to install the [ LabVIEW TensoRT toolkit ].

1. Introduction to LabVIEW graphical TensorRT toolkit

Toolkit Features:

1. Graphical programming : You can complete machine vision projects without mastering the basics of text programming;
2. Direct model conversion : Onnx model (part) can be converted to FP32, FP16 or Int8 tensorRT model (.trt or .engine);
3. Extremely fast inference interface : load the tensorRT model and perform extremely fast inference (the speed is 2~5 times that of Onnx-tensorRT);
4. Custom layer network : For experienced players, you can use the INetworkDefinition advanced tool to create your own network, view or edit existing Onnx networks;

Performance evaluation comparison of yolov5s under various tool packages:

注：测试电脑cpu为i7-11800H，Intel集显为i7-11650G7，独显为笔记本RTX 3070（包括预处理和后处理）

The running speed of yolov5s under the tensoRT toolkit:

2. Installation of LabVIEW graphical TensorRT toolkit

1. Preparations before installation

• Make sure your computer hasNvidia discrete graphics card, if not, it is recommended to use other toolkits. The TensorRT toolkit requires the support of Nvidia independent graphics card;
• Before installing the LabVIEW graphical TensorRT toolkit, please install LabVIEW 2018 64-bit or higher version . Remember to require the 64-bit version . If it is already installed, there is no need to install it again. Regarding the installation of LabVIEW, there are many tutorials on the Internet, so I won’t go into details here;
• In order to successfully install the toolkit, it is recommended that you upgrade VIPM to VIPM-21 . If there is a higher version or this version is already installed, there is no need to install it again; VIPM-21 download link: https://download.csdn.net/download/ virobotics/85463218
• Install the AI ​​vision toolkit: techforce_lib_opencv_cpu-1.0.0.98.vip. The installation steps can be viewed at: https://blog.csdn.net/virobotics/article/details/123656523

2. LabVIEW graphical TensorRT toolkit and related driver installation and configuration

Configure cuda and cudnn

• ① If your computer has an Nvidia independent graphics card, configure cuda 11.3 and cudnn (the following examples will use 11.3 as an example).
• ② The installation package provides an installation-free package for cuda+cudnn. The installation-free package path is: "...\cuda113\bin.7z". Unzip the bin compressed package and place it in any location. Set the bin folder inside as an environment variable.
  
• ③ Execute nvcc –V in the command line to check whether cuda has been successfully configured. The following shows successful configuration.
  
• ④ Shut down and restart the computer;

Install the TensorRT toolkit

1. Make sure that any version of cuda 11.0~11.6 (11.3 is recommended) and the corresponding cudnn have been installed, and the environment variables have been configured;

2. Before installation , please set labview.exe to be started by the administrator , as shown below (please remember to change it back after the installation is complete)
   

3. Open vipm as administrator

4. Double-click and run [ virobotics_lib_tensorrt-1.0.0.22.vip ] to start the installation. The installation process will automatically install the tensoRT driver package;

5. Wait a few seconds and the following interface will appear. Click Install to install;

   

6. The installation takes a few minutes, wait for a while, and the following interface will appear, all showing No Errors, which means the installation is successful, click Finish;
   

7. After successful installation, restart LabVIEW and create a new VI. Right-click on the block diagram panel (remember it is the block diagram panel, not the front panel ) –> click Addons –> you can see that there is an additional item "VIRobotics" in the additional tool package Addons – >Click VIRobotics–>Click function selection version TensoRT to find the related functions of TensoRT in the tool package we just installed. Drag into onnx_to_engine.vi. If it can run, the installation is successful.
   

3. Common installation errors

1. The VIPM installation interface shows that it is always connected to LabVIEW. After opening LabVIEW, you see a message indicating that the environment variable has failed to be added.

Solution: Click OK to continue installing the toolkit. After the installation is complete, manually add the following path to the system environment variable Path.

C:\ProgramData\VIRobotics\driver\TensorRT\lib

2. Drag onnx_to_engine.vi to the block diagram panel and report an error

Solution: Please make sure that the computer's cuda has been configured, and make sure that \TensorRT\lib has been added to the system environment variables. If an error is still reported, it is recommended to shut down and restart.

3. Download the LabVIEW graphical TensorRT toolkit and related drivers

If you would like a tool kit, please leave your email address in the comments section

Summarize

The above is what I want to share with you today. If you have any questions, you can discuss them in the comment area. If you think the content is good, you can like it and save it. If you want to discuss more about LabVIEW and artificial intelligence technology, welcome to join our technical exchange group. Please note CSDN when joining the group

If the article is helpful to you, please follow, like, and collect it