Where are 2019 Fastai resources to use home workstation/GPU?

Dear Experts,

I’ve now spent a number of months sourcing and configuring
hardware to make a machine learning workstation.
(LOTS of surprise factor.)

The fastai V1 software stack is also installed.

conda list fastai

packages in environment at /home/ml22/anaconda3:

Name Version Build Channel

fastai 1.0.61 1 fastai

The 2020 fastai course is not completed yet.
So, I plan to start the 2019 fastai course.

The forums say that the new material
won’t work on the old software stack

“fastai v2 and the new course were released on August 21st.
fastai v2 is not API-compatible with fastai v1”

And, this weekend, the website was changed.

The new link for the 2019 is:

I know I need at least:
jupyter notebooks (code)
data, and maybe more

Now that things have moved, what are the
correct links for the 2019 course?

What is the procedure to use them?


What exactly is unclear and what resources do you need? Everything should be available on the https://course19.fast.ai
If you already have fastai v1 installed, then you should be able to just continue with the lessons.

But I would recommend overall to instead work on the newest course because there is no active support for fastai v1, and the new course contains the same material plus more.

Yes I know, -somewhere- under

“I know I need at least:
jupyter notebooks (code)
data, and maybe more”

To download everything to my server,
What/Where are the exact links?

Ie. Lesson 1
This link has the video

But what are the links to download the
Jupyter notebook?


In the video links on the course website, you should see a sidebar on the right with notes. If you scroll down, for each lesson it should have the notebooks. Here is the github repo with the notebooks as well:

BTW if you haven’t started the course yet, I’d recommend moving to the 2020 version of the course instead.

Thanks a lot!

So that’s where they were hiding.

2019 course, Lesson 1 notebook is at:


I was able to download the notebook, and
run it on my workstation/GPU

My concern was:
where/how do I download the data, and get the
notebook, python, and data to work together?

Turns out that the notebook downloads the
datasets as it goes.

I was not expecting that. If I thought that
would happen, I thought there would be a
wget with an exact full URL,
but didn’t see either.


Thanks again!