I meet a problem when using tensorRT on successive frames of video without a “sleep”. The program works normally at first executions, but much slower in later ones. When I put an interval of several ms between 2 inputs, every execution consomes the “normal” time.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| TensorRT frames processing speed increases with increase in number of frames | 4 | 832 | October 15, 2021 | |
| why not accelerate use tensorrt? | 1 | 366 | March 13, 2020 | |
| I found that using tensorrt for inference takes more time than using tensorflow directly on GPU | 1 | 747 | April 9, 2019 | |
| TensorRT 7 quit without any error? | 1 | 609 | March 2, 2020 | |
| Ideas to maximize throughput using TensorRT | 1 | 365 | November 20, 2020 | |
| Slow first inference and very slow two models inference | 3 | 1250 | August 2, 2022 | |
| tensorRT5 inference speed slown down in multithread application. | 2 | 1323 | December 3, 2019 | |
| Why is TensorRT faster than TensorFlow? | 3 | 1660 | April 26, 2022 | |
| Getting stuck while running Jetson Nano Tensort inference | 3 | 603 | September 4, 2019 | |
| Inference doesn't speed up if not continuous after initialization | 3 | 533 | September 15, 2020 |