Lesson 1: fastai course on Azure DSVM

Dear helpful reader,

I am trying to setup fastai course on a rental GPU Deep Learning Server as per this suggestion / instruction in Lesson 1:

Getting a GPU Deep Learning Server

To do nearly everything in this book, you’ll need access to a computer with an NVIDIA GPU (unfortunately other brands of GPU are not fully supported by the main deep learning libraries). However, we don’t recommend you buy one; in fact, even if you already have one, we don’t suggest you use it just yet! Setting up a computer takes time and energy, and you want all your energy to focus on deep learning right now. Therefore, we instead suggest you rent access to a computer that already has everything you need preinstalled and ready to go. Costs can be as little as US$0.25 per hour while you’re using it, and some options are even free.


I have followed the instructions for:
Azure Data Science Virtual Machine Azure Data Science Virtual Machine | Practical Deep Learning for Coders (fast.ai)

but no joy…

I have figured out that the recommended VM NC-series have been retired as on 6 Sep 2023 (oops) and hence the script:
bash fastai2onAzureSpotDSVM.sh → no longer works.

I have tried following the instructions for Google Cloud Platform Google Cloud Platform | Practical Deep Learning for Coders (fast.ai) with no success either.

I have googled for solutions to no avail.
And searched youtube for video solutions to no avail.
I have searched this forum and reached the 'Azure" topic and clicked onto the pinned resources and followed the instructions therein, but no success yet…


  1. Is setting up fastai course jupyter notebook on a rental Full Linux Server a worthwhile educational activity?
  2. If 'Yes", would any kind soul walk me through setting up the fastai course jupyter notebook on either Google Cloud Platform or Azure Data Science Virtual Machine? Or please point me to a online resource on achieving this?

That sounds like the current state when the book was published, however the current V1 course has everything in Kaggle / or Colab to save the hassle of setting up an environment. (The book material is still used as reference, though the order and some lesson content has changed as you would expect in a fast moving field.)

There is no need to go down the renting a Full Linux Server yet, and when that is something, you wish to pursue (ie because you’ve hit the limits of those services in size or speed requirements), then Jeremy’s live code sessions are a great help there. Live-coding (aka walk-thrus) ✅

Thanks Allen for your advice. Your advice is greatly appreciated : )