Lesson 1 - Official topic

The official release date is July 14th, but with the pandemic, it might be pushed further…

8 Likes

Can we have a signed version of the book from you guys?

14 Likes

Oreilly’s official release date is June 2020.

2 Likes

Bastille day ! :slight_smile:

3 Likes

When will the MOOC for the Data Ethics class be released? I really wanted to go to the in-person class but couldn’t this time around.

7 Likes

This may not be something that you want to address, but I’m curious if the funds from the book go to the fastai team or to O’Reilly.

1 Like

We haven’t set a definitive date yet for releasing the Data Ethics MOOC, but it will be July at the latest.

16 Likes

I don’t want to share a -ve emotion, but I want to say this out loud again, if you feel intimidated by “CS Background” people having a “better start” in this amazing field.

I studied CS at University-IMO, it didn’t make me a better coder. Practising code made me a better coder.

Please don’t let that intimidate you. :slight_smile:

17 Likes

To piggy-back on that, I have little education on CS and only knew Windows before taking the fastai course two years ago. You can learn the “geeky/hacker” things as you go, you don’t need to know any of them beforehand.

26 Likes

i studied management and learned how to code through fastai. its definitely possible to start from 0

11 Likes

Just want to second the opinion - I didn;t know how to code when I started fast.ai last year for the first time and eventually learnt my way bit by bit. It’s absolutely not necessary to know “everything” about Python at the start, but only a basic understanding of what’s going around is really helpful.

Everything can else be learnt day by day through continuous effort.

6 Likes

“… without any human training or control” . Isn’t the training part of building a DNN a “human training”? We use human-generated labels after all

Yes, in that regard, it is under human control (and this part is crucial for ethics). What that statement means is that a human is not coding step by step how the algorithm is making its decision.

3 Likes

I wrote more about the fast.ai teaching philosophy in this post:

12 Likes

“Play the game first” is exactly why I love Fast.ai and prefer it over any other source of ML studying.

3 Likes

It would be great to apply LEGO SERIOUS Play as part of this Whole Game as well. Awesome idea Rachel and all

I want to share my writeup on “How not to do fast.ai (Or Any ML MOOC)”

TL;DR:
Don’t take a bottom up approach, follow the Top down approach like Jeremy just mentioned :slight_smile:

5 Likes

The credit goes to David Perkins, Paul Lockhart, and others who inspired us.

1 Like

Is it easy to convert Torch to PyTorch? What are the differences?

Will we be learning about how to put Pytorch models into production? Or will this primarily be for one off projects?

2 Likes