My First Steps post concluded with a wish to “run into more problems in the space of coding and implementing neural network architectures”. I did succeed in running into problems… …on my way to getting started with coding and implementing neural network architectures.
I wanted to “simply” feed some other data into the convolutional neural network from this tutorial. In order to do so, I had to get that data onto my Azure Virtual Machine (VM). The data I wanted to try was too large for my existing VM, so I needed to add additional disk storage. This proved to be easier said than done. After a number of more courageous than competent
sudo commands, I had succeeded in breaking my VM. By “pizza hour” (4 pm on Tuesday), I was truly starting to despair of the situation. Fortunately, my fellow scholar Alethea, who has much more experience than I do in software engineering and systems administration, came to my rescue. In lieu of a longer blog post about my learning this week, here is a writeup of how to mount extra storage onto your Azure VM, should you ever run into the same conundrum as I did.
Next week, I plan to get to the point of a minimum viable network for classifying neuroscience data… …hopefully with fewer logistical blockers.