DS5.0 BuildingTensorRT for objectDetector_Yolo takes 2.5 mins - object detector yolo

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hi, when running the objectDetector_Yolo app it takes 2.5mins to build the TensorRT engine, although I already have a model-engine-file. The 2.5 mins starts when the INFO [TRT]: has just shown all the layers e.g. conv_1, leaky_1 etc. It takes 1.5 mins to come back with the next line (INFO [TRT]: Some tactics do not have sufficient memory (etc) and then a further 1 min to end with ‘Building complete’. Any advice would be appreciated. Thanks

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) Xavier NX
• DeepStream Version 5.0
• JetPack Version (valid for Jetson only) Most recent
• TensorRT Version 7.1.0

2

The config file needs to reference the .engine rather than the .weights.

#model-file=yolov3.weights
model-engine-file=yolov3_b1_gpu0_int8.engine