[DRIVE AGX]I' want to use nvidia-smi

Hello,
I need nvidia-smi to use tensorflow.

because this question has not been resolved.
https://forums.developer.nvidia.com/t/tensorrt-fails-for-custom-ssd-inception-model/68049

How can I use the nvidia-smi?
Thanks.

Hello @Masayuki.Kamoda,
can you please provide the following information:
Hardware Platform: [Example: DRIVE AGX Xavier™ Developer Kit, DRIVE AGX Pegasus™ Developer Kit]
Software Version: [Example: DRIVE Software 10,]
Host Machine Version: [Example: native Ubuntu 18.04]
SDK Manager Version: [Example: 1.0.1.5538]

also, can you please elaborate on your use case? (in addition to the referenced post)

thanks

Dear shayNV

I want to run an object recognition engine of our original design.

Hardware Platform: DRIVE AGX Xavier™ Developer Kit
Software Version: DRIVE Software 10
Host Machine Version: docker in native Ubuntu 18.04
SDK Manager Version: sdkmanager_1.0.1-5538_amd64

Thanks

Dear @Masayuki.Kamoda,

nvidia-smi is our tool for non-embedded platforms, our alternative tool on embedded platforms is tegrastats.
you can read some information on the tool here

if you need to profile your application please use the NSIGHT SYSTEMS /NSIGHT COMPUTE tools delivered with the installation through SDK Manager.

If you are not planning to use one of the built-in NN of DW that is delivered with the SDK, and you have your own NN for object classification please consider our webinar for usage:
Integrating DNN Inference into Autonomous Vehicle Applications with NVIDIA DriveWorks SDK

1 Like

Dear shayNV,

Thank you for sharing your webinar with me.

Is only Caffe or ONNX supported to run DNN with the current version of DriveAGX?
In that case we will have to change the framework from TensorFlow.

thanks

Hi @Masayuki.Kamoda,

In the current release Drive Software 10, also the uff model format is supported to be converted to TensorRT plan file.
but please notice that for next release and future releases too this option will be deprecated.

Shay