some newbie questions :)

Hello!

I recently purchased a jetson TX2 . I am currently a third year student of computer engineering. I have been using tensorflow for quite sometime now and I bought Jetson to use it as a cheap alternative to train my models on. I have a bunch of questions and I know these my be really dumb questions but I have no idea where to start … I would be really grateful if someone could answer me. :

  1. I have never owned any SPC , I was wondering if someone can share a book or any other type of guide for a complete newbie so that I can get started with this.

After my first impression, I can tell that this is an amazing piece of hardware. I was wondering what else can be done on this ? is there a list of sample projects that can be made using jetson?

  1. I have followed JetsonHacks to build tensorflow on jetson but wasn’t quite able to due so which is due to an update in java software if I am not mistaken. due you guys know any other guides on how to build tennsorflow on jetson .

here is the link to JetsonHacks tensorflow build repo on github for your references :

  1. Is there anyways that I can install Miniconda , Berryconda or any other Conda like python package manager on Jetson ? I did my best to find answers on the internet … it seems like other people have experienced the same issue but no one had a proper answer … I would appreciate it if you can point me to the right direction .
    Also, is there anyway that I can install Jupyter Notebook ?

Here is a list of posts I found on this topic :

https://devtalk.nvidia.com/default/topic/901150/python-anaconda-on-tx1/

https://devtalk.nvidia.com/default/topic/1002785/jetson-tx2/anaconda-or-miniconda-on-jetson-tx2-/

  1. I saw someone building a cluster of Jetson TX1 … I was wondering if there are any guides to do the same for Jetson TX2 ?

I have seen others build clusters of Rasberry Pi but those have much much lower performance compared with Jetson . While it seems like great fun , I was wondering about real life applications of a TX1 or TX2 cluster considering that these are significantly more powerful than any other SBC on the market.

here is a link for your references to TX1 cluster :

https://devtalk.nvidia.com/default/topic/761234/jetson-tk1/jetson-tk1-mpi-cluster/post/4267732/#4267732

Bump

Hi damoon.azarpazhooh, you may be interested in our Two Days to a Demo deep learning tutorial for Jetson here:

https://developer.nvidia.com/embedded/twodaystoademo

There are also additional tutorials available here: https://developer.nvidia.com/embedded/learn/tutorials
And a list of projects & guides on the wiki here: https://elinux.org/Jetson_TX2#Guides_and_Tutorials

You may be interested in this thread, where pip wheel for TF 1.5 is shared:

https://devtalk.nvidia.com/default/topic/1028840/jetson-tx2/available-tensorflow-1-5-for-jetson-tx2/
https://github.com/JesperChristensen89/TensorFlow-Jetson-TX2

Conda, I am not aware of a redistributable produced by the maintainers for aarch64/arm64, hence it isn’t typically available.

Yes, you should be able to install ipynb/jupyter from the apt repo, I’ve run that before on Jetson, seems to work normally.

The Jetson TX2 and TX2 devkits and modules are essentially backwards-compatible from a hardware perspective, their devkits are basically the same, so you should be able to follow the same guide for the most part.