Share this page!

François Chollet Profile picture
Twitter logo
14 Oct, 4 tweets, 2 min read
The results of the global survey of professional ML developers and data scientists (run by Kaggle) just dropped! Many interesting insights, including fresh data about ML framework adoption.
kaggle.com/kaggle-survey-…
Perplexing that Keras dropped by 3 pts compared to last year while TensorFlow increased by 3 pts, since 99% of new TF usage leverages Keras. I think this illustrates the fact that the TF brand tends to override the Keras brand, i.e. many of our users answer "TF" but not "Keras".
A similar phenomenon was likely at work in the StackOverflow developer survey 2021, which showed a large adoption gap between TensorFlow and Keras -- which should not exist in 2021...
Keras usage has meaningfully grown in 2021 in absolute numbers (e.g. docs traffic), though that doesn't say anything about market share (for which we have to look at other metrics, including these 2 surveys). TF/Keras share is unchanged compared to 2020 as far as I can tell

• • •

Missing some Tweet in this thread? You can try to force a refresh
　

Keep Current with François Chollet

François Chollet Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @fchollet

François Chollet Profile picture
10 Oct
Harry Potter and the Philosopher's Stone sold 120M copies, The Little Prince sold 100M. 7M over 64 years is tiny! A big difference is that these books had global audiences, but Ayn Rand is US-only. Hardly anyone in Europe or Japan has even heard of her.
To everyone saying "actually...": I can't speak for the UK or India, but in non-English speaking countries she is unknown. Atlas Shrugged didn't have a French edition until 2011, a German edition until 2012, Japanese until 2004. It sold *thousands* of copies in these countries.
Meanwhile in the US she's completely mainstream. Everyone knows her name and the title of that book. Very commonly read in high school for class. So yes, there's a big difference. The extent she's known elsewhere is a pure side effect of her influence in the US.
Read 4 tweets
François Chollet Profile picture
9 Oct
Do you ever need to get actual users and use cases if you can just keep churning out hyped-up catchphrases, hashtags, and Twitter threads instead? I guess we'll find out eventually.
I've seen a number of hype waves over time -- some that I jumped on, some that flopped, some that worked out. Web 2.0, smartphones, SoLoMo, 3D printers, VR, deep learning, chatbots, wearables... one thing is sure, "web3" has the lowest content-to-hype ratio of anything I've seen
Short list of things that I was enthusiastic about very early, basically on day one: web 2.0, smartphones, DL (back when all the ML folks thought DL was a fad). Things I was cautiously optimistic about: SoLoMo, VR, early crypto (2012-2013).
Read 6 tweets
François Chollet Profile picture
9 Oct
Lots of folks (e.g. in my replies) confuse being able to name something and being able to explain it (i.e. understand it).

"Why does each half of the earth get cooler/hotter in opposite yearly cycles?"
"Duh, that's just seasons"
The ultimate version of this being extremely hot takes like "that's just brain development". I mean, sure. Also, computing is just electrons and art is just pixels. Easy. I have all the answers. I know the words!
Sometimes it's a bit subtle.

"How do airplanes fly?"
"That's just *lift*"

"Why does a child born from parents with brown and blue eyes respectively have brown eyes?"
"That's just because the *allele* for blue eyes is *recessive*"
Read 4 tweets
François Chollet Profile picture
30 Sep
A portfolio project for undergrads? There's a 300 lines of code keras.io example that does exactly this (with better performance). Anyone with some Python experience could read it and modify it.

Machine learning just keeps getting easier and more accessible.
One thing I want to emphasize: ease-of-use is obviously beneficial to those with less expertise, who go from 0 to 1. But the people who gain the most from it are actually the experts, who are now able to dream bigger and move faster. Experts go from 1 to 10.
"My expertise is devalued, undergrads can reimplement my PhD thesis in 30 minutes 😢" is the wrong mindset.

It's actually: "my expertise is being multiplied, I can now achieve dramatically more ambitious milestones while leveraging more enjoyable and productive workflows 🚀🚀🚀"
Read 4 tweets
François Chollet Profile picture
30 Sep
Here's a trivial example to illustrate the difference between pattern recognition and reasoning and it impacts behavior generation: let's say you encounter, for the very first time, a glass door with ⅃⅃Uꟼ written on it.
Pattern recognition: nearest neighbor is "PULL", I've learned to associate that with pulling the door. I pull.

Reasoning: that's a mirror image of "PULL". Must be written on the other side. The door will open by pulling towards the other side, i.e. pushing from this side. I push
typo: "and *how* it impacts behavior generation"
Read 4 tweets
François Chollet Profile picture
19 Sep
Some apparently read this as "all recommender tech is bad & dangerous" which is not it at all.

The takeaway is that the information you consume matters and your attention is precious, so you should have very high standards for what's allowed to download thoughts into your brain.
You wouldn't eat random crap handed to you by a greasy robot on the street. Being deliberate about what you eat is critical to your health.

Well, being deliberate about the information you consume is critical to your mental well-being.
That doesn't mean that you should not eat anything or that you should never go to restaurants. It just mean you should have standards and you should be mindful of what you eat.
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!

Like this author's thread?

Get emails when @fchollet has new unrolls!

Check out other threads from @fchollet!

Read all threads by @fchollet

:(