Custom Library in Graph Composer

Please provide complete information as applicable to your setup.

• Hardware Platform GPU
• DeepStream Version 6.0.1
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
**• Issue Type questions

Hi there,
I’m trying to use graph composer.
In deepstream-python, I used a custom tensorrt inference (e.g. yolov5) and add LD_PRELOAD before everything.

However, I have no clue to do the same thing in composer. The app reports errors without the .so file which is generated while building tensorrt .engine file.

Any ideas?

thx for the help.

I make it work by using command lines. Add LD_PRELOAD before /opt/nvidia/graph-composer/execute_graph.sh.
However, I still want to know if there is another solution?

Your solution is right.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.