V2 web site online, launching soon

I’ve put the new course up on course.fast.ai , and moved the old course to course17.fast.ai . I’ll be doing a launch announcement in a few hours. In the meantime, could you let me know if you come across any problems on either site, or have any suggestions for improvements?

Also if you have more detailed contributions, pull requests to the web site github are also welcomed! :slight_smile: https://github.com/fastai/dlcert2


Lesson 6 Video is wrongly linked to lesson2?

Thanks - well spotted!

Also please check lesson 5 video. Its linked to lesson 6.

emdash didn’t render on the front page:

And there’s a whole community of thousands of other learners ready to help you with your journey &emdash; just head over to forums.fast.ai if you need any help, or just want to chat to other deep learning learners.

“Theano” and “keras” show up on http://course.fast.ai/about.html as opposed to PyTorch and the fastai library.

http://wiki.fast.ai/index.php/Python_libraries is out of date (theano, keras) and linked in http://course.fast.ai/about.html

Thanks @Matthew they’re fixed now.

BTW adding stuff to the FAQ would also be appreciated! http://forums.fast.ai/t/part-1-faq

I may sound a bit picky :innocent: but after looking quickly at the home of http://course.fast.ai/, and clicking on the tabs (Home, About, Getting Started, etc.), I couldn’t find an “obvious” link about “Installing the Fast.ai library, required for the course”.

Also, when it comes to installing the Fast.ai library (as it often pops up on KaggleNoobs), I am myself not capable of giving a clear answer with regards to the instructions on GitHub: which is it best to use today ?

  • First, clone fastai using git, Then, cd to the fastai folder and create the python environment: approach or
  • Go straight to pip install fastai


Thanks for the suggestions - picky is good! The recommended approach is conda env update for the course. The pip installer is only updated occasionally, FYI.

The first lesson links to @reshama’s paperspace installation instructions FYI. http://course.fast.ai/lessons/lesson1.html . Installing on your personal machine isn’t recommended except for particularly motivated and tenacious students who need an excuse for why they bought an expensive GPU box :wink: .

1 Like

Got it !

My take is, remembering my noobish path with a gaming GPU: if I read the Fastai GitHub page today, I would go straight for the easiest path of pip install fastai, not the conda env update one with 5+ instructions to enter, because it is not obvious that “The pip installer is only updated occasionally, FYI.”.

I realize I’m being a very picky beta-tester here but “been there, done that”, I’m the annoying fellow in your focus group :sunglasses:

That’s a good point. I just updated the readme.

1 Like

Maybe a note for university students that AWS gives a few credits under the ‘AWS Educate’ pack?
Or (They offer 300$ as of now)…

I found these useful given my financial constraints :sweat_smile:, I think other college students might find these useful too.

Maybe you could add that to the FAQ under a question like “How can I access a GPU for minimal cost?”

Added the question under category ‘Cloud Compute Options’

1 Like

Sorry if I’m missing an obvious link, but I can’t find how to get to the Part 2 notebooks. (I had everything over on Amazon, but I’m trying to switch to Paperspace. From the git pull, I only get dl1 and ml1). Thanks!

The 2017 repo is still at https://github.com/fastai/courses/ , which is where you’ll find part 2.


@jeremy Your deep learning course is great! I want to take your Deep learning part II in Spring 2018. I applied through website, but haven’t received confirmation. Could you tell if there is something wrong?