To return to your notebook, the basic steps will be:

  1. Start your instance
  2. Update the course repo
  3. Update the fastai library
  4. When done, shut down your instance

Step by step guide

Start your instance

In your console notebooks choose the notebook you want to run and click on the button ‘Start’ under action.

You can choose a different Virtual Machine type on which you’d like to run your Notebook. This can be extremely useful when you want to start on a lower-end machine type, test everything is okay, then move to a more powerful GPU. Also, sometimes the GPU type that you started the notebook on will be unavailable, in which case you can easily fire it up on a different GPU.

Click on ‘Start notebook’ when you’re ready and wait a few seconds for it to be ready.

Click on the open button when it’s ready and you’ll be back in your jupyter notebook page.

Update the course repo

To update the course repo, launch a new terminal from the jupyter notebook menu.

This will open a new window, in which you should run those two instructions:

cd course-v3
git pull

This should give you the latest of the course notebooks. If you modified some of the notebooks in course-v3/nbs directly, GitHub will probably throw you an error. You should type git stash to remove your local changes. Remember you should always work on a copy of the lesson notebooks.

Update the fastai library

To update the fastai library, open the terminal like before and type

pip install fastai --upgrade 

Stop your instance

Once you’re finished, go back to your console notebook page and find your running notebook. Under Action, click stop, this will end the billing session.

NOTE: you will be charged for the time that your notebook is running. You must stop the notebook to stop incurring charges.