Possibility of using python and/or Tensorflow on DRIVE AGX

Hi,

We have some models trained with Tensorflow Object Detection framework and want to test it on DRIVE AGX. I wonder if there are any tutorials of running python scripts on DRIVE AGX. I understand cross-compilation with C/C++ is the official approach, but adopting the existing python script tested on Jetson Xavier could be a quick start for us. Thanks for the explanation.

Hi,

Do you want to deploy the model with TensorRT?
Currently, TensorRT python API doesn’t support DRIVE AGX so it’s recommended to try the c/c++ approach.

If you want to run the model with TensorFlow frameworks directly, it should be workable.
You will need to build TensorFlow from source and run the python script as the way in other environment.
Here is topic for building TensorFlow on DRIVE for your refererence:
https://devtalk.nvidia.com/default/topic/1049100/general/tensorflow-installation-on-drive-px2-/post/5324624/#5324624

Thanks.

Thanks, the link is extactly what I need. I will study how to convert from tf to tensorRT model.