Lesson '0': Practical Deep Learning for Coders (fast.ai)

00:00 - Two groups of students in general
02:01 - The fastai book
02:55 - The course
03:22 - Finish the course!
04:27 - Finish a project!
05:49 - What can a project be?
06:53 - Be tenacious!
08:27 - Radek Osmulski story
10:15 - Stop endlessly preparing for doing deep learning
11:16 - What will fastai teach you
12:25 - How to get started with coding
14:00 - The missing semester of your CS education
15:30 - Share your work or learning
17:25 - four steps to do fastai lessons
20:32 - Notebook Server vs Linux Server
23:43 - Get started with Colab
29:37 - Github with Colab
30:37 - Clean version of notebook
31:26 - Questionnaires
32:32 - Share your model on your dataset
34:29 - Wrong ways to do fastai
36:41 - Start positive learning feedback
37:27 - Read and Write code
38:07 - Immerse yourself in DL world through twitter
40:39 - Go blogging
42:03 - A great thing to blog
44:03 - How ML differs from other coding
45:15 - Why and How to create a good validation set
46:20 - Coding DL is harder than other forms of coding
47:16 - Baseline for your project
49:51 - Kaggle competitions as best projects
52:33 - Build your portfolio for job
55:30 - Get to be the firsts to do part 2
56:08 - How to get started with AWS EC2

[Finn-jppn]

I remember being so excited to learn how to train my own models. I was following through with all your sessions back in 2019. But then life happened and I couldn't catch up. And ever since I felt like I missed to a train that I could've boarded. Seeing this video today has awoken that same old excitement within me. I will try my best to follow through this time and complete the whole series. Thank you, Jeremy ❤️

[anangelsdiaries]

I love how Jeremy just dropped "17 years-old PHD graduate" as if it was the most normal thing. What the heck.

[MaxwellMcKinnon]

V4!!!! Woot woot. Thanks Jeremy for enabling me to get in the door on ML at Apple and now Google, having an awesome time making crazy stuff. I gave up too easily on the PhD required tag at one point trying to self study my way in. I was very happy i stumbled onto fastai (back when it was v1 keras). I remember it being so hard to get started ~5 years ago, even as a Math lover and math major, most resources were way too pre-req math focused and not enough applied and intuition building focused. Love your dive in and go (Mathematician’s Lament) approach. Sort of like JIT compilation: don’t exhaustively study the pre-req but rather study the end product first then use that context to motivate the study of the needed parts of the pre-reqs ;). Love and much thanks!!! (Admittedly I didn’t finish every fastai lesson, but it’s because my projects took off, and at that point doing fastai became the getting ducks in a row activity :D)

[petergoodall]

With respect to keeping it simple…

“There are two ways of constructing a software design: One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. The first method is far more difficult.”


― C. A. R. Hoare

[hamedgholami]

every year better than last year. Jeremy, you inspire me to be a good person and help people have better lives. thanks for setting a really good example.

[oleksandrdovgusha]

That is the video for me. Thanks, Jeremy! I am the guy who actually has some domain knowledge but postpones the practical application and spends more time on another math book/course.

[abhishekaiem]

Oh Man, I had just started and completed 3 lesson in that... again will start from begining, anyway these sessions are gold which can be watched several times. Thanks Jeremy :)

[arenashawn]

This is an amazingly inspiring video and I know I tend to dive into things head first and been really pumped up to start, and I tend to read by consuming a lot of content because I have very good memory. But then at some point I crash and never finish things. I want to stick to to this one and really finish this journey. Hope to see you all on the other side!❤❤❤

[DavidPunsalan]

The clean versions of the notebooks is awesome idea. It is amazing how much thought has been put into this!

[khairulumam]

You cares a lot about your community :')

thank you so much for your patience and your effort on this community

Glad to know fast ai

[pariscatblue]

Thank you Jeremy for this video, and very much! I'm one of those that started and never finished... and not because I didn't want to...
And I have vastly benefited from that part that I managed to complete... whatever, here I am again and will do my best to finish it this time!
:-)

[CihatUysal]

Thank you Jeremy! Super exited for part 2.

[SimplyAndy]

Why not add this to the lecture website. Maybe as an optional vid. This is really helpful, but many learners might be unaware of this vid.

[datasciyinfo]

Thank you so much for fast.ai v4. Our Meetup is using your course to learn together, in Virginia. We’re waiting for Part 2, having finished most of Part 1. Also Radek Osmulski sounds just like me, except I already went down that math rabbit hole for a different subject, so now I don’t have that distraction. Still taking a long time to learn ml and dl, about 3 years for me so far. Just starting to feel I know the very easy basic things. Python libraries and coding tools were most time consuming and difficult to remember for a non-coder.

[michaeltranchina]

I am assuming that most of what is in this fantastic video is now (2024) obsolete...I am afraid to ask if there will be a new updated video put together to bring everything up to date?

[techworld]

amazing I love this ... I follow all the series now

[adamcosgrove]

If possible, please can you create a video from start to finish on setting up and finding the API key discussed in lesson 2 of the 2020 course? I really enjoy the course but am stunted by my inability to be able confidently to find it. And, all solutions on the forums appear to be outdated with people commenting on similar issues. Thanks

[datasciyinfo]

I rewatched this video, and I’d like to add that Part 2 is Much Much Harder than Part 1. I’ve done 4 chapters or so from Part 2, without writing my own versions of fastai library (further research topics). I only try to understand how fastai parts work to apply it to my own project. Diving into fastai code is complex. Uses advanced coding concepts. I am slowly understanding the small parts of fastai library needed for my project, but I totally understand why only a few people will finish Part 2.

[thryce]

i would like to recommend "A Matrix Algebra Approach to Artificial Intelligence " by Xian-Da Zhang. I never actually read it haha but the preface is one of the sweetest things ever. The author was in uni in china during the cultural revolution. Ended up being called burgouis and sent to be reeducated on a farm for a decade or so on what it meant to be "chinese" He eventually was able to immigrate to the US and in his mid 30s worked and went and got his phd (i think by hisearly 40s) . HIs family wrote that instead of spending his final time with them he pushed thru to complete that book because he thought it was imp to pass down knowledge to the next generation so they wouldnt repeat the mistakes of the past. its a very academic text and seems rather dry but nice to read his story and to remember that others have been thru more shit than you and persevered. and if you dont finish this its probably not because of circumstance because but because you made a decision that other things were more imp. totally ok ofc but at least be honest about it

[Teslawaverunner]

Will there be a new Part 1 i.e. will there be a “Deep learning for coders 2021 “ ? Or is it just a Part 2 that we can expect this year ? Thx