how to install anaconda, pytorch, tensorflow, numpy and opencv on px 2

Autonomous Vehicles DRIVE PX 2 General
1

hi,

i want to set a python 3 environment on my px 2. and i really dont know how to install anaconda, pytorch, tensorflow, numpy and opencv on a aarch64…

thanks all…

2

Dear hanbinl,

There is no anaconda aach64 file to install. You might have to get support from the anaconda community for this.
OpenCV is not an official feature, if you want to build opencv check the below links
[url]https://devtalk.nvidia.com/default/topic/1019937/faq/opencv/[/url]
https://devtalk.nvidia.com/default/topic/1032172/faq/drivepx2-opencv3-cuda9-patch/provide instructions
For Tensorflow, please follow the thread [url]https://devtalk.nvidia.com/default/topic/1049100/general/tensorflow-installation-on-drive-px2-/[/url].

We have TensorRT which can optimize the models generated using pytorch on host, can be deployed on Drive PX2 for inferencing. Could you please check pytorch sample at [url]https://docs.nvidia.com/deeplearning/sdk/tensorrt-sample-support-guide/index.html#network_api_pytorch_mnist[/url]

3

hi SivaRamaKrishna,

thanks for your help, i have another question, should i install cudnn for cuda 9.2 on taget and host. i really want to know how to build a perfect environment in the beginning.

Also, if i have a very clean ubuntu 16.04, what should i do in steps on host and taget…

thanks again
hanbinl

4

Dear hanbinl,
When you flash your board using SDKManager, it installs CUDA and cudnn libs on host and target. Please check the instructions in https://developer.nvidia.com/DRIVE/secure/files/SDKManagerUserGuide.pdf to flash the host and target