I would like to introduce tensorRT into the existing yolov3 object detection system of the company

When I tried object detection with jetson nano, the fps was slow, so I want to use tensorRT to speed it up. I tried demos using tensorRT. I tried this.
https://jkjung-avt.github.io/tensorrt-yolov3/
After trying the demo, I tried to integrate it into an existing company’s object detection system, but I’m not sure how. The existing system uses darknet_ros. If you have any advice, please let me know. Thanks!

Hi @kubo1,
Do you mean you want to optimize your custom yolo model using TensorRT?
The below link should be able to help you in that case
https://jkjung-avt.github.io/trt-yolov3-custom/

Also you can check the link


For queries related to darknet_ros, you can raise it here.
Thanks!

Oh, yes! The information I wanted to know.
Thanks!!