Print average model inference time using tensorrt executable

elight1 · November 3, 2022, 3:57pm

Description

I have noticed that with the latest tensorrt.exe version which was recently released (8.5), converting an onnx model using the terminal ‘tensorrt’ command, does not show the average/mean inference time at the end like before.

Is there a command line argument which prints them? I have checked everything in ‘tensorrt --help’ but nothing would pring the average inference time…

spolisetty · November 7, 2022, 6:17am

Hi,

Please try trtexec --verbose, which provides the summary along with TensorRT engine building.

Thank you.

system · December 7, 2022, 12:14pm

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

Topic		Replies	Views
Could not infer onnx model for TensorrtExecutionProvider provider TensorRT tensorrt , onnx	1	1169	November 11, 2022
Tensorrt Execution Provider TensorRT tensorrt , cudnn , onnx	1	821	November 27, 2023
TensorRT inference take too much time than expected TensorRT tensorrt	2	1032	December 22, 2020
Inference TensorRT randomly returns nan TensorRT tensorrt	2	538	April 27, 2023
Inference time of tensorrt 6.3 is slower than tensorrt 6.0 TensorRT tensorrt , driveos	7	916	October 12, 2021
Different engines give different inference results when using the same onnx model and giving the same input TensorRT	4	957	December 31, 2023
What is meaning of time output from TensorRT? TensorRT	1	402	August 25, 2021
TensorRT Engine gives incorrect inference output for segmentation model TensorRT	6	1322	October 12, 2021
TensorRT 8 : C++ inference gives different results compared to tensorflow python inference TensorRT	7	1359	October 5, 2021
ONNX runtime prediction using GPU and with different intervals TensorRT	4	2009	January 19, 2022

Print average model inference time using tensorrt executable

Description

Related topics