[Deepstream 3.0] Floating point exception when running with nv-tracker

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hi,
I am trying to run YOLOv3 with tracker in Deepstream 3.0. The way I am doing it right now is by creating TensorRT engine of YOLOv3 and add it to primary inference module.

deepstream-app runs smoothly with only inference module. When I add nv-tracker to the pipeline, it ran and showed result for a few minutes, then threw floating point error and core dump. However, the time when the error appears is different, from 10-15 minutes to even an hour. Below is the error log

**PERF: 25.53 (25.03)
**PERF: 6.67 (25.02)
**PERF: 42.03 (25.03)
**PERF: 11.06 (25.02)
**PERF: 44.58 (25.03)
**PERF: 27.04 (25.03)
**PERF: 25.10 (25.03)
Floating point exception (core dumped)

Any help is appreciated.

2

HI, Anhqle
Have you resolved this issue? Can you reference deepstream-plugins/sources/apps/deepstream-yolo and add nvtracker to it then try again?

3

Anhqle

Can you refer to https://github.com/vat-nvidia/deepstream-plugins
deepstream-plugins/sources/apps/deepstream-yolo
And add nvtracker to the pipeline.