I’m just venting here, but I’ve spent the last 5 hours trying to get my machine back after bricking it from trying to install Linux. And that’s after wasting another 6-8 hours yesterday getting the GPU to work on my windows machine only to find out it can’t get past the training in the cats vs dogs problem.
I followed the instructions for Windows 10 and got Lesson 1 up to the point where it started to train and ended up crashing in the middle.
And I googled the error message I got and it advises installing Linux
So I figured it shouldn’t be too bad since I’ve been installing Linux for dual booting on many machine for over a decade… little did I know that since the new machine I’m using right now uses UFEI, I’m in for a world of pain…
I burned Linux Mint onto a USB stick and my machine won’t boot it. I googled why and it seemed I needed to disable UFEI. So I did. And the installer loaded. Then it was time to partition my drive and I set aside some free space for a linux partition. And proceeded to install Linux.
Then when I rebooted, I was greeted with Linux. Great. But I still want to use Windows for everything besides Fast.ai. And I never got that Grub or whatever window upon rebooting my machine.
Here’s what I’ve tried so far to restore my Windows partition:
Booting into Linux and running ‘sudo update-grub’: it doesn’t detect my windows partition.
Running ‘sudo os-prober’ : still doesn’t find my windows partition
Creating a Windows 10 boot disk on a USB drive. I’ve wasted the past 3 hours doing this. Since I couldn’t boot from windows, I had to download and burn the USB boot disk from Linux and every time, it takes like 30 minutes. First burn was a failure. 2nd time was also a failure. Upon booting, my machine completely skipped the USB drive. I tried turning UFEI back on again and my machine refused to boot at all. At least with UFEI turned off it boots into Linux.
Yesterday, I wasted like 6-8 hours trying to get fast.ai working with GPU support on my Windows 10 machine (granted, several months ago, I had gotten Tensorflow to work fine with GPU support with ZERO hassles but for some reason fast.ai didn’t like my CUDA installation and I had to upgrade everything to CUDA 9 and TF 1.5 in all of my Anaconda environments.) And today just as I thought I could finally be doing some exercises, my machine is now unable to boot back into Windows and another day is wasted repairing it (and still no end is in sight.)
I should have gone with the cloud option from the start… sigh…