Implement YOLO on DRIVE AGX Xavier

[JaeJ]

Please provide the following info (check/uncheck the boxes after creating this topic):
Software Version
DRIVE OS Linux 5.2.6
DRIVE OS Linux 5.2.6 and DriveWorks 4.0
DRIVE OS Linux 5.2.0
DRIVE OS Linux 5.2.0 and DriveWorks 3.5
[O] NVIDIA DRIVE™ Software 10.0 (Linux)
NVIDIA DRIVE™ Software 9.0 (Linux)
other DRIVE OS version
other

Target Operating System
[O] Linux
QNX
other

Hardware Platform
NVIDIA DRIVE™ AGX Xavier DevKit (E3550)
[O] NVIDIA DRIVE™ AGX Pegasus DevKit (E3550)
other

SDK Manager Version
1.9.1.10844
[O] other

Host Machine Version
[O] native Ubuntu 18.04
other

I have a YOLO PyTorch model.
Basically, it was written as Python script.
To implement this YOLO on the DRIVE Xavier, Should i convert it to CUDA C++ and compile to use GPU computation?
Also , i try to use GPU directly in docker container but it doesn’t work well.

[SivaRamaKrishnaNV]

Dear @JaeJ
Docker is not supported on DRIVE AGX Xavier platform.
For yolo model, convert Pytorch → ONNX and then use TRT c++ APIs on target for ONNX-> TRT conversion and perform inference.

[JaeJ]

Thanks. I’ll try to do that :)

[JaeJ]

Also you mean, to implement YOLO, i should convert all the python code to CUDA C++?

[VickNV]

Please refer to Yolov5 model on Drive AGX.

[JaeJ]

Thanks

[SivaRamaKrishnaNV]

@JaeJ

to implement YOLO, i should convert all the python code to CUDA C++?

No need. Just prepare a YOLO ONNX model to start using TRT C++ APIs.

[JaeJ]

I see!