Welcome to the 2022 fast.ai course!

If you are reading this, it means you are part of the 2022 fast.ai course - welcome! The course will cover material from the first half of the book, plus new material (such as the Transformers architecture for natural language processing). You strongly recommend prerequisites of a year of coding experience, and high school mathematics.

There are two groups that will be attending:

  • People who are part of the official University of Queensland course
  • Selected members of the fast.ai community that have been selected for the fellowship program

The first lesson will be April 26th, 6pm, Brisbane Australia time zone. Online streaming will be available via YouTube. Recordings will be available immediately for everyone in this group, and will be available later this year to the general public. Click on the following links to see the lesson times in your time zone:

Be sure to visit this thread for more information: FAQ, resources, and official course updates :white_check_mark:. And head over here to introduce yourself: Introduction thread.

Don’t hesitate to reply below if you have any questions.


Thank you, @jeremy for giving us so much! I am eagerly looking forward to the course, and especially the dissection of the Transformer architecture that you’ll do!


Superb, soo happy to receive this invite. Eagerly waiting for the lessons.

1 Like

Oh wow, that was a long time since I participated in the course! And many changes happened in the Deep Learning landscape. Would be great to know Jeremy‘s opinion on the recent trends, like multi-billion parameter LMs and Transformers in general. The fast.ai courses was the first place where I learned about applying transfer knowledge for NLP. So maybe we’ll get some great insights this time as well :slight_smile:


WOW! thanks a lot Geremy, for this new course and for sending me an invite, I am so very happy about this invite.

Infinite Thanks and blessings for all you give Geremy.


Thank you @jeremy for welcoming us! Looking forward to learning more stuff.

Is it already cool/acceptable to publish new topics in this 2022 forum about anything related to fastai?

Sure feel free. :slight_smile: Best to use this thread unless it’s an issue large enough to need its own thread: General course chat


Correction: I wrote originally “Note that in the US this will be April 25th.” But that’s wrong. It’s actually first thing in the morning April 26th.


I’m so happy to hear that your courses are staring up again. They’ve already made an incredible difference in my life. I hope you’ll let us know how you’ve been doing these last couple of years and how your plans to boot up Deep Learning in Australia are going. Thanks very much for inviting us! :smiley:


I’m so very happy to be here! I’m looking forward to the course and I’m reading the book again to be better prepared. Beyond the content, I’m also eager to enjoy those practical pearls of advise that are dropped casually during the classes, and to reconnect with this fantastic community. Thank you!


Let’s go!

Glad to have the course back and glad to be back with the community to enjoy it. A couple of things …

First, for folks new and old, I’m blogging the relevant part 1 portions of the fastbook at ohmeow.com (and yes, I will be done before lesson 1).

Second, for folks looking to get a jumpstart with training transformers with fastai, I just released version 1 of my blurr library and also posted a blurrified version of one of Jeremy’s notebooks that shows how to use it for a regression problem.

See y’all on the forums, discord, twitter, etc… etc…


Thank you so much @jeremy, happy to receive an invite.
Looking forward to learning new stuff!!

1 Like

Thanks a lot for inviting me @jeremy . Can’t wait to learn more on the transformers, as I know you always have amazing insights to give :slight_smile:

1 Like

This is the only course I take every time it comes out. Even if it starts from the basics: there’s always something new, Jeremy & Rachel have this independent way of researching that’s just :man_cook:t2::pinched_fingers:, and this community is the best and most accessible way to get into ML.

I started here years ago and learned ML hand in hand w software dev. Now coming from industry I’m looking forward to seeing what’s new to learn and helping out.

Thanks a lot for the invite, I really appreciate it.


Thank you so much for the updated course. Looking forward to learn and contribute.

1 Like

Glad to hear the course is offered again and thanks for the invite. @jeremy your teaching is incredible and fast.ai is generous to offer such learning opportunities for everyone. Eager to learn new material on Transformers from you as I tend to suggest this to one of my PhD student who is trying to develop an AI model to generate Python code from clinical documents.


Thanks for the invite Jeremy. Feels good to be back :slight_smile:

1 Like

I’m looking forward to it. :+1:

wonderful!, it’s always a great day for the AI community when fast.ai releases something, looking forwards :slight_smile: