Deepstream-app rebuilds TensorRT engine for custom Yolo model on each run

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hi,

I followed the guide for integrating custom Yolo model

Everything works perfectly except the point that the TensorRT engine is rebuilt each run.
Could you please give me a hint how to deal with it?

3

Hi,

Please update the model engine path so Deepstream can start from deserializing instead.

Ex. configs/model_yolo.txt
Please make sure below filename correct and the engine exists.
Sometimes it fails from serialization due to the permission of the folder.


[property]
...
model-engine-file=model_b1_gpu0_fp16.engine
...

Thanks.

5

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