Can I run this yolov5 in Python samples?

[OctaM]

I am using Python sample apps and I want to use this model. Do I have to follow the exact same steps?

1. Converting the model to onnx
2. Compile the plugin and deepstream parser?

[Tyler-D]

If you only want to run python sample, you don't need the decode plugin.

Follow the steps of TensorRT sample:
https://github.com/NVIDIA-AI-IOT/yolov5_gpu_optimization#tensorrt-sample

Just comment this line to avoid the plugin loading:
https://github.com/NVIDIA-AI-IOT/yolov5_gpu_optimization/blob/main/tensorrt-sample/yolov5_trt_inference.py#L16

[OctaM]

Thanks for the answer @Tyler-D.

What about those samples? https://github.com/NVIDIA-AI-IOT/deepstream_python_apps

Can I run it inside those? And if yes, should I also remove the decode plugin?

I've been trying to run it inside an app with this config https://github.com/NVIDIA-AI-IOT/yolov5_gpu_optimization/blob/main/deepstream-sample/config/config_infer_primary_yoloV5.txt.

I also compiled the plugin and generated yolov5_decode.so but when I run the sample it doesn't show any error just Segmentation fault (core dumped).

[OctaM]

Update: I was able to do this by following https://github.com/marcoslucianops/DeepStream-Yolo