Skip to content
By using Twitter’s services you agree to our Cookies Use. We and our partners operate globally and use cookies, including for analytics, personalisation, and ads.
  • Home Home Home, current page.
  • About

Saved searches

  • Remove
  • In this conversation
    Verified accountProtected Tweets @
Suggested users
  • Verified accountProtected Tweets @
  • Verified accountProtected Tweets @
  • Language: English
    • Bahasa Indonesia
    • Bahasa Melayu
    • Català
    • Čeština
    • Dansk
    • Deutsch
    • English UK
    • Español
    • Filipino
    • Français
    • Hrvatski
    • Italiano
    • Magyar
    • Nederlands
    • Norsk
    • Polski
    • Português
    • Română
    • Slovenčina
    • Suomi
    • Svenska
    • Tiếng Việt
    • Türkçe
    • Ελληνικά
    • Български език
    • Русский
    • Српски
    • Українська мова
    • עִבְרִית
    • العربية
    • فارسی
    • मराठी
    • हिन्दी
    • বাংলা
    • ગુજરાતી
    • தமிழ்
    • ಕನ್ನಡ
    • ภาษาไทย
    • 한국어
    • 日本語
    • 简体中文
    • 繁體中文
  • Have an account? Log in
    Have an account?
    · Forgot password?
    New to Twitter?
    Sign up
fchollet's profile
François Chollet
François Chollet
François Chollet
Verified account
@fchollet

Tweets

François CholletVerified account

@fchollet

Deep learning @google. Creator of Keras. Author of 'Deep Learning with Python'. Opinions are my own.

United States
fchollet.com
Joined August 2009

Tweets

  • © 2021 Twitter
  • About
  • Help Center
  • Terms
  • Privacy policy
  • Cookies
  • Ads info
Dismiss
Previous
Next

Go to a person's profile

Saved searches

  • Remove
  • In this conversation
    Verified accountProtected Tweets @
Suggested users
  • Verified accountProtected Tweets @
  • Verified accountProtected Tweets @

Promote this Tweet

Block

  • Tweet with a location

    You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more

    Your lists

    Create a new list

    Under 100 characters, optional
    Privacy

    Copy link to Tweet

    Embed this Tweet

    Embed this Video

    Add this Tweet to your website by copying the code below. Learn more

    Add this video to your website by copying the code below. Learn more

    Hmm, there was a problem reaching the server.

    By embedding Twitter content in your website or app, you are agreeing to the Twitter Developer Agreement and Developer Policy.

    Preview

    Why you're seeing this ad

    Log in to Twitter

    · Forgot password?
    Don't have an account? Sign up »

    Sign up for Twitter

    Not on Twitter? Sign up, tune into the things you care about, and get updates as they happen.

    Sign up
    Have an account? Log in »

    Two-way (sending and receiving) short codes:

    Country Code For customers of
    United States 40404 (any)
    Canada 21212 (any)
    United Kingdom 86444 Vodafone, Orange, 3, O2
    Brazil 40404 Nextel, TIM
    Haiti 40404 Digicel, Voila
    Ireland 51210 Vodafone, O2
    India 53000 Bharti Airtel, Videocon, Reliance
    Indonesia 89887 AXIS, 3, Telkomsel, Indosat, XL Axiata
    Italy 4880804 Wind
    3424486444 Vodafone
    » See SMS short codes for other countries

    Confirmation

     

    Welcome home!

    This timeline is where you’ll spend most of your time, getting instant updates about what matters to you.

    Tweets not working for you?

    Hover over the profile pic and click the Following button to unfollow any account.

    Say a lot with a little

    When you see a Tweet you love, tap the heart — it lets the person who wrote it know you shared the love.

    Spread the word

    The fastest way to share someone else’s Tweet with your followers is with a Retweet. Tap the icon to send it instantly.

    Join the conversation

    Add your thoughts about any Tweet with a Reply. Find a topic you’re passionate about, and jump right in.

    Learn the latest

    Get instant insight into what people are talking about now.

    Get more of what you love

    Follow more accounts to get instant updates about topics you care about.

    Find what's happening

    See the latest conversations about any topic instantly.

    Never miss a Moment

    Catch up instantly on the best stories happening as they unfold.

    1. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020

      in general, many simple cs101 homework code-gen "success-stories", but also many many failures. at least it knows that it should add some tests:pic.twitter.com/tMVq21xTM7

      3 replies 0 retweets 3 likes
      Show this thread
    2. François Chollet‏Verified account @fchollet 20 Jul 2020
      Replying to @yoavgo

      What if you prompt an alternative description of fizzbuzz with a rules twist? Does it generate classic fizzbuzz or does it adapt?

      2 replies 0 retweets 3 likes
    3. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
      Replying to @fchollet

      what kind of alternative did you have in mind? (one issue is that prompting with such variations may change the task from doing "fizz buzz" to doing "how to change my prompt examples for this new setup". which is fair, but need to be aware of this.)

      2 replies 0 retweets 0 likes
    4. François Chollet‏Verified account @fchollet 20 Jul 2020
      Replying to @yoavgo

      Write a program that prints all numbers from 100 to 1000. But for numbers that are not multiples of five print "Fizz" instead of the number and for numbers that are not multiples of seven print "Buzz". For numbers which are multiples of neither five nor seven print "FizzBuzz".

      1 reply 0 retweets 2 likes
    5. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
      Replying to @fchollet

      ah, this is a much lower level of abstraction then what i asked for. ok, will test.

      1 reply 0 retweets 2 likes
    6. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
      Replying to @yoavgo @fchollet

      first, it clearly saw such variations in its training data:pic.twitter.com/lbwSRt7OKx

      1 reply 0 retweets 0 likes
    7. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
      Replying to @yoavgo @fchollet

      second, its results vary in quality, but i am pretty sure that with enough tries it will get it right.pic.twitter.com/vVJNZxtPfd

      1 reply 0 retweets 0 likes
    8. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
      Replying to @yoavgo @fchollet

      oh, i missed your negation there. hmm, maybe it won't get it.

      1 reply 0 retweets 0 likes
    9. François Chollet‏Verified account @fchollet 20 Jul 2020
      Replying to @yoavgo

      Yes, the negation matters. The test here is whether it reads the text and adapts to it, despite looking a lot like something it has fully memorized -- arguably you fail that test, too ;-)

      1 reply 0 retweets 1 like
    10. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
      Replying to @fchollet

      :) yeah i definitely patterned-match a lot when reading it.

      2 replies 0 retweets 1 like
      François Chollet‏Verified account @fchollet 20 Jul 2020
      Replying to @yoavgo

      What I've been trying to determine is whether the few-shot capability is just "text completion scaled up" or wether it is qualitatively different from previous capabilities. So far it seems like the former.

      10:47 AM - 20 Jul 2020
      • 7 Likes
      • Arip Asadulaev Pranjal Yadav Erfan Noury Robert Oschler Arpit Gupta eder deliprao.eth (📜,📜)
      2 replies 0 retweets 7 likes
        1. deliprao.eth ( 📜, 📜)‏ @deliprao 20 Jul 2020
          Replying to @fchollet @yoavgo

          deliprao.eth ( 📜, 📜) Retweeted deliprao.eth ( 📜, 📜)

          This is my thinking too, especially around my experiences with scaling up BERT models for factoid QA type completions. I have a suspicion it’s a mixture of the two cases you mention. https://twitter.com/deliprao/status/1284631048572121088?s=21 …https://twitter.com/deliprao/status/1284631048572121088 …

          deliprao.eth ( 📜, 📜) added,

          deliprao.eth ( 📜, 📜) @deliprao
          Replying to @yoavgo @earnmyturns @minimaxir
          a lot of the zero-shotness comes from the amount of data used in training. We see this even at a smaller scale BERT models. And when Fernando is saying "IR", it's not just smoothing over words -- it's more like retrieval and smoothing over entire parse (sub)trees.
          0 replies 0 retweets 2 likes
          Thanks. Twitter will use this to make your timeline better. Undo
          Undo
        2. New conversation
        1. (((ل()(ل() 'yoav)))) 👾‏ @yoavgo 20 Jul 2020
          Replying to @fchollet

          i think it is a combination of both. obviously it did not learn to code from instructions or to understand short textual stories, this is just "text completion scaled up". but it *did* learn to pick up on non-trivial task definitions from 1-2 Q/A examples, and carry them through.

          2 replies 0 retweets 1 like
        2. Prannay Khosla‏ @PrannayKhosla 20 Jul 2020
          Replying to @yoavgo @fchollet

          @sibeshkar

          0 replies 0 retweets 0 likes
          3. End of conversation

      Loading seems to be taking a while.

      Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

        Promoted Tweet

        false

        • © 2021 Twitter
        • About
        • Help Center
        • Terms
        • Privacy policy
        • Cookies
        • Ads info