Getting low fps (object detection with yolov8s) on Jetson Xavier NX

Autonomous Machines Jetson & Embedded Systems Jetson Xavier NX
, , , ,
1

Device : Jetson Xavier NX 16GB Ram
Tensorrt Version : 8.2.1.9
jetpack version : 4.6.1
model conversion : pytorch → onnx → engine(using trtexec)
when inference this engine on video or camera it gives only 12-15 fps
How can I improve FPS

3

Hi,

Have you maximized the device’s performance?

$ sudo nvpmodel -m 0
$ sudo jetson_clocks

Thanks.

4

how can i use this

5

Simply run the above two commands on your shell, then run your applications.

6

there is no improvement in FPS after running these commands

7

Is Fan on the heatsink running?

8

yes

9

how can i increase fps

10

Hi,

Could you check the GPU utilization ratio that reports by tegrastats?

$ sudo tegrastats

If TensorRT already fully utilizes 99% of resources, you will need another device that has the more computational power to increase fps.

Thanks.

12

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