yolov5s fp16 reasoning is slower than fp32 opened 08:18AM - 28 Jul 23 UTC (UTC) [QZ-cmd] QZ-cmd question ### Search before…

Jetson agx orin executes yolov5s bug

AastaLLL July 31, 2023, 6:16am 5

Hi,

Have you maximized the device’s performance?

$ sudo nvpmodel -m 0
$ sudo jetson_clocks

More, do you have the ONNX model?
Could you check to get the benchmark data with trtexec?

Thanks.

Topic		Replies	Views
I encouter some problems when I run yolov5 with AGX Orin Jetson AGX Orin cuda , python	2	31	November 22, 2024
Improving the speed for fp32 for yolov10x inference from Ultralytics on Jetson AGX Orin 64g devkit Jetson AGX Orin tensorrt , camera , yolo , python , tensorrt-model-optimizer	5	82	September 18, 2024
YOLOv5 performance on Jetson Xavier AGX Jetson AGX Xavier camera , yolo , pytorch	2	2164	October 18, 2021
Inference slow even using TensorRT Jetson AGX Orin tensorrt	15	1890	November 6, 2023
Why the number of flops is different between FP32 and FP16 mode with YOLOv3 TensorRT implementation? Jetson AGX Xavier tensorrt , kernel , profiling	8	4054	October 18, 2021
YOLOv5S model performance testing benchmark DeepStream SDK jetson , deepstream	3	56	April 25, 2025
xavier ternsorrt mnist fp16 is slower the fp32? Jetson AGX Xavier	10	875	October 18, 2021
Slow inference with yolov8 pytorch on agx orin Jetson AGX Orin yolo	7	2133	September 5, 2023
Why I get much higher TFLOPS in Orin AGX than what claimed in the document IGX Developer Kit kernel , jetson-inference , documentation	7	351	November 4, 2024
Why jetson nano fp16 is slower than fp32 Jetson Nano jetson-inference	2	629	October 15, 2021