Deep Learning Examples for Drive AGX devices

Autonomous Vehicles DRIVE AGX Orin DRIVE AGX Orin General
1

Please provide the following info (tick the boxes after creating this topic):
Software Version
[*] DRIVE OS 6.0.10.0
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other

Target Operating System
[*] Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
[*] DRIVE AGX Orin Developer Kit (940-63710-0010-100)
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
other

SDK Manager Version
2.1.0
other

Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
[*] native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
other

Issue Description
I am looking for some sample solutions to execute DNN primitives like object detection and semantic segmentation using the repository

I did the necessary compilation environment and when I try to execute one of the sample application from build directory :

jetson-inference/build/aarch64/bin/imagenet.py images/orange_0.jpg images/test/output_0.jpg

I get the error message as shown below :
image

I am aware that Nvidia provides sample applications for object detection & other vision primitives but I am looking for python based solution so I referred to above repository. Can you please let me know how this error can be solved or is there any repository available similar to jetson-inference for Drive AGX series?

2

Dear @PA_GN ,
Jetson inference repository is for Jetson devices and not expected to work with DRIVE.
We don’t have python based E2E sample for DRIVE.
As you aware, DW provides an example to show object detection. You may refer to TensorRT python samples and check implementing model as a sample.
Also see GitHub - NVIDIA/DL4AGX: Deep Learning tools and applications for NVIDIA AGX platforms. if it helps for your case.

3

Hi @SivaRamaKrishnaNV,

I am actually trying to execute semantic segmentation for lane detection as mentioned in the Jetson Inference repository but want to execute on Drive AGX Orin. Is there any similar repository or reference available for semantic segmentation (lane detection).
For object detection we have a python based solution which uses TensorRT.

4

Hi @SivaRamaKrishnaNV,

I checked the link GitHub - NVIDIA/DL4AGX: Deep Learning tools and applications for NVIDIA AGX platforms and want to use UniAD-TensorRT solution but it mentions that to run inference application an model engine needs to be built which requires Model Training & Data Preparation. I checked the link of Nuscenes but the data size is more than 250 GB. Is there any pretrained model file available which can be used for validation?

5

Dear @PA_GN ,
Due to licence issue, we could not provide the pretrained model in the repo.

8

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