Getting Ubuntu and my GTX 1060 Setup for TensorFlow

Things on the TensorFlow installation front did not go as planned yesterday. The Ubuntu 16.04.1 installation worked at first. I was able to grab an ISO and use Rufus to load it on a USB drive. Things loaded without any hiccups, but my GeForce GTX 1060 graphics card was not recognized. I tried several different methods to get it to work. All of them resulted in what some folks in the online forums called an NVidia login loop. After the PPA and driver installation Ubuntu would load, but nothing happened after successfully logging into the system. I gave up on installing Ubuntu 16.04.1 and moved to working with the Ubuntu Studio flavor. That change in course made all the difference.

Installation Steps Taken

1. Downloaded Ubuntu Studio 16.04.1 ISO using my Windows 10 base machine
2. Dropped the ISO on a USB via Rufus
3. Setup drivers for Geforce GTX 1060 via sudo add-apt-repository ppa:graphics-drivers/ppa
4. Ran sudo apt-get update && sudo apt-get distro-upgrade
5. Rebooted the entire computer
6. Logged back into the system and opened additional drivers
7. Picked 367.44 from Nvidia-367
8. Clicked apply changes
9. Rebooted
10. loaded with no Nvidia login loop 😉

My next course of action was to get Google Chrome installed from https://www.google.com/chrome/browser/desktop/index.html.

Leave a Reply