Here’s a quick update of FPS numbers (on Jetson Nano) after I updated my tensorrt yolov4 implementation with a “yolo_layer” plugin. For example, “yolov4-416” (FP16) has been improved to 4.62 FPS. Refer to my jkjung-avt/tensorrt_demos repository for details.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Scaled-YOLOv4 in Deepstream | 5 | 1062 | October 12, 2021 | |
| Jetson nano Not able to give higher FPS | 8 | 2205 | October 15, 2021 | |
| Full Yolov3 on the nano using TensorRT or Deepstream 4.0.1 | 7 | 2496 | October 14, 2021 | |
| Low fps when doing object detection on jetson nano | 19 | 8858 | March 1, 2022 | |
| Yolov4 tiny in DeepStream 5.0 | 8 | 4289 | October 12, 2021 | |
| Yolov3 is very slow | 21 | 20207 | October 14, 2021 | |
| run yolov3-tiny with tensorRT model | 7 | 3391 | January 4, 2020 | |
| deepstream-yolo-app performance vs Tensor-Core optimized yolo-darknet | 9 | 3621 | October 12, 2021 | |
| Yolov3 in nanojetson | 12 | 1074 | October 18, 2021 | |
| Options for optimising custom TFRT tiny YOLOv4 implementation to improve live inference speed on Nano | 6 | 1080 | October 15, 2021 |