Now that @YouTube has decided, in their infinite wisdom, to remove dislike count, making it much harder for you good folks to figure out which of my videos are popular, I figure I'm gonna have to go through them and pick out the top ones for you.
A 🧵 of my most liked vids: 1/
A 🧵 of my most liked vids: 1/
I'll pick those with >100 likes, with min 99.4% "like" to "dislike" ratio.
#1, at 100% like ratio, is my explanation of exactly how the fastai library works, and *why* it works that way, presented to the Swift4TensorFlow dev team.
youtu.be
#1, at 100% like ratio, is my explanation of exactly how the fastai library works, and *why* it works that way, presented to the Swift4TensorFlow dev team.
youtu.be
#2, at 99.6% likes, is my talk about what I learned in my time building startups. I put a lot into developing this talk and am really proud of it, so I'm thrilled it's been such a hit with you!
youtu.be
youtu.be
The next three are all equal, with 99.4% likes.
This talk, which I presented to an academic conference, describes why I don't think tools like Python, PyTorch, and Tensorflow are the long-time future of deep learning, and what I hope for the future:
youtu.be
This talk, which I presented to an academic conference, describes why I don't think tools like Python, PyTorch, and Tensorflow are the long-time future of deep learning, and what I hope for the future:
youtu.be
In this talk, I discuss how I did stuff before deep learning came along... and why, once it did arrive, I switched nearly entirely to using deep learning for most of my modelling.
youtu.be
youtu.be
And in this talk I give my tips on how to succeed at @fastdotai. It draws heavily from the book "Meta Learning" by @NVIDIAAI deep learning expert @radekosmulski.
youtu.be
youtu.be
And finally, a couple that don't exactly qualify, but I'm including anyway...
This video, from way back in 2008, is my appearance on Australia's largest TV morning news program, talking about IE's latest security fail.
It's got 100% likes, but a low count since no-one followed me back then.
youtu.be
It's got 100% likes, but a low count since no-one followed me back then.
youtu.be
Lesson 12 of my Intro Machine Learning course has 100% likes, but that's kinda cheating because if you've watch all the way to lesson 12, you clearly like the material!
It's a full walk-thru of a complex modeling task, plus a discussion of data ethics.
youtu.be
It's a full walk-thru of a complex modeling task, plus a discussion of data ethics.
youtu.be
(There's quite a few lessons with 100% likes (or thereabouts) but I won't include them all here since, as I say, it's cheating :D )
I've only picked videos I've done that are on my YouTube channel. There's a bunch more around the place that aren't on my channel. If there are any other videos I'm in that you think are worth watching, pop them in the replies!
One interview that I've heard a lot of folks say they like is my chat with @lexfridman -- he's a really thoughtful interviewer who does his homework better than nearly anyone else I've spoken to:
youtu.be
youtu.be
BTW the rankings above are based on like-ratio. Based on views, my most popular videos, as you'd expect, are lesson 1 of each of my main courses. Plus (surprisingly, to me) the March-2020 videos we did raising the idea that masks are perhaps a good idea.
All those videos are well-liked too, but only Practical Deep Learning for Coders Lesson 1 has broken the 99% barrier!
Loading suggestions...