hello thanks you for your support at advance,
i’m new in ML and i was satified by 5 fps using yolov3 tensorrt till i found this link( How to use GPU + 2 DLA can be 100FPS for YoloV3 on Xavier ) that talk about the performance of xavier nx with yolov3 model it reached 50fps, i tired to reproduce it but i couldn’t, if somebody could help me doing this or explain to me this result i will be appreciated. thanks an other time
Hi,
The performance is measured by TensorRT in INT8 mode.
You can reproduce it with the source below:
In case you don’t know.
You can maximize the device performance with the below command:
$ sudo nvpmodel -m 0
$ sudo jetson_clocks
Thanks.
The script you gave me is for benchmark right ? How can i get this fps with inference pls i’m confused for darknet i get 2 fps using tensorrt i get 5 fps and i want something real time like 30 fps for example thanks you
The script you gave me is for benchmark right ? How can i get this fps with inference pls i’m confused for darknet i get 2 fps using tensorrt i get 5 fps and i want something real time like 30 fps for example thanks you
Hi,
Have you maximized the device performance?
For darkent Yolo, please check our Deepstream SDK for inference.
It has optimized the whole pipeline on Jetson.
/opt/nvidia/deepstream/deepstream-6.0/sources/objectDetector_Yolo/
Thanks.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.