segmentation fault （core dumped）

[ghty11468]

The engine file has been generated, but this error always occurs during inference. What should I do?

[marcoslucianops]

Please follow: https://github.com/marcoslucianops/DeepStream-Yolo/blob/master/docs/YOLOv5.md

The engine output is wrong, so you are using the wrong ONNX model.

[ghty11468]

请关注： https://github.com/marcoslucianops/DeepStream-Yolo/blob/master/docs/YOLOv5.md

引擎输出错误，因此您使用了错误的 ONNX 模型。

My onnx file was generated using the Ultratics/Yolov5 project, but is it not allowed?

[marcoslucianops]

You need to use the export_yoloV5.py file from the utils folder in this repo to export the ONNX model from the YOLOv5 PyTorch model (as said in the link I sent you, please follow it).

[emad555]

You need to use the export_yoloV5.py file from the utils folder in this repo to export the ONNX model from the YOLOv5 PyTorch model (as said in the link I sent you, please follow it).

This does not generate an engine file, only onnx file.

[marcoslucianops]

The engine file is generated in the first time you run the deepstream-app command.

[emad555]

Thank you, it is working now!