Update - Thanks to James post below, I learned/realized that the dl2 course notebooks currently live under
and not alongside the first course:
where I was expecting it (as dl2)…since that’s where dl1 lives.
hence wasting a lot of time with git when in fact I already had the notebooks, but didn’t realize it.
It seems I must be overlooking something but after spending 20 minutes, time to ask here:
What is the proper git command to sync and get the notebooks for the part 2 course?
I can see the notebooks of course here:
But can’t seem to be able to get a local copy to thus work with.
*I’ve updated fastai, tried to clone, etc. the notebooks but they refuse to pull down for me. I also pip installed the dev version ala:
pip install “git+https://github.com/fastai/fastai#egg=fastai[dev]”
I mean that after running a relevant command (which I’ll list below), my fastai folder has nothing contained for the “course-v3/nbs/dl2”…I still have “course-v3/nbs/dl1” but no notebooks for part 2.
This should get you going. I’d suggest deleting your local copy of fastai_docs or, in the event that you have some local changes there that you’d like to keep, you can just move the repo to a temporary location outside of where you do the following. Also note that the part II is a separate repository from part I. Part I is in the fastai/course-v3 and Part II is in fastai/fastai_docs repo (most likely because it just started and its development is still in progress).
Make sure you aren’t in another repository. i.e. you shouldn’t be in course-v3 at all. Note: My setup on most machines is I have a develop folder in my home directory that all my projects/repos live.
The standard way is to use the “fork-and-branch” workflow. It’s a bit complicated because it allows to manage several use-cases and probably overkill for this course notebooks. There’s a simpler workflow that may work for you.
The general idea is to pull the last version of the notebooks from the official fast.ai repository but push your changes to a remote repository of your own.
In your local repository you will have two branches: master and your own (for example: ‘MY_WORK’). On master there will be the clean, updated fast.ai version, on ‘MY_WORK’ your work along with the fast.ai changes.
First create an empty repository in your Github. For instance, ‘MY_FASTAI_DOCS’ (you can mark it private if you like.)
Here’s what worked for me:
Forcing myself to add versioning to my little experiments. I’d usually find a way around it because of my laziness but really using it day to day will make you find out all of the fine details and how for ex: Diff branches work, how would you submit a PR to the fastai repo, etc.
Kickstarting a repo may seem overkill at first, but it pays off quickly. You’ll be able to test different approaches easily, go back and forth the changes, review them while keeping a local and a cloud backup. Besides with GitHub you can review your notebooks online without having to start Jupyter and/or the dev environment.
hey, thanks for asking this question! thank god! I was looking for the nb too and wasnt able to find it. I thought I was bit “shy” to ask this question since it just seems that everyone can just figure it out on their own.