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. Petr Baudis‏ @xpasky 30 Oct 2017

      Petr Baudis Retweeted Rossum

      @fchollet check out @bzamecnik 's perf. analysis of #Keras multi-GPU training speedups - many detailed measurements https://github.com/rossumai/keras-multi-gpu/blob/master/blog/docs/measurements.md …https://twitter.com/RossumAi/status/924985495851012096 …

      Petr Baudis added,

      Rossum @RossumAi
      Can you train big #Keras neural nets on many GPUs? Apparently we are one of the first to try! We're solving this now https://medium.com/rossum/towards-efficient-multi-gpu-training-in-keras-with-tensorflow-8a0091074fb2 …
      1 reply 0 retweets 0 likes
    2. François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @xpasky @bzamecnik

      I did my own benchmarks with Xception and ResNet50 (nothing fancy) and got 85% efficiency with 8 GPUs on EC2 (6.8x speedup)

      1 reply 0 retweets 2 likes
    3. François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @fchollet @xpasky @bzamecnik

      When training from an ImageDataGenerator, performance was worse (65% or so efficiency). Potentially because of slow disk I/O on EC2

      1 reply 0 retweets 1 like
    4. François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @fchollet @xpasky @bzamecnik

      Adrian got 97% efficiency on a GoogleNet-like with 4 GPUs (74 min to 19 min): https://www.pyimagesearch.com/2017/10/30/how-to-multi-gpu-training-with-keras-python-and-deep-learning/ … -with 4 GPUs, I had low 90s efficiency

      1 reply 3 retweets 5 likes
    5. François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @fchollet @xpasky @bzamecnik

      Overall I think our parallelism model is sound (per-device forward and backwards pass on sub-batches + CPU concat, keep params on CPU)

      1 reply 0 retweets 1 like
    6. François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @fchollet @xpasky @bzamecnik

      But it would need auditing, in particular to check whether we correctly place the grads computation on each device (now we leave it to TF)

      1 reply 0 retweets 1 like
    7. Petr Baudis‏ @xpasky 30 Oct 2017
      Replying to @fchollet @bzamecnik

      His NVIDIA DevBox is PCIe 16x, Azure is PCIe 8x, custom builds often even worse. Scaling also impacted by small batch sizes (64 -> 32).

      1 reply 0 retweets 0 likes
    8. Petr Baudis‏ @xpasky 30 Oct 2017
      Replying to @xpasky @fchollet @bzamecnik

      Overally we focus on small batch sizes, as I see huge sample sizes in many real world applications, including internally at @RossumAi.

      2 replies 0 retweets 0 likes
    9. François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @xpasky @bzamecnik @RossumAi

      We have significant overhead (shuffling the params from CPU to GPU at every step; split+concat). Need time(process(sub_batch)) >> overhead

      1 reply 0 retweets 1 like
      François Chollet‏Verified account @fchollet 30 Oct 2017
      Replying to @fchollet @xpasky and

      Not doable to reduce the overhead, so to have a speedup you need to keep your per-sub-batch processing time high (large models or batches)

      9:39 AM - 30 Oct 2017
      • 1 Like
      • Evgeni Arent
      1 reply 0 retweets 1 like
        1. New conversation
        1. Petr Baudis‏ @xpasky 30 Oct 2017
          Replying to @fchollet @bzamecnik @RossumAi

          We have ideas to reduce the overhead - https://github.com/rossumai/keras-multi-gpu/blob/master/blog/docs/conclusion.md … Most of it is the samples, first step: StagingArea https://gist.github.com/bzamecnik/f76e480edf98e95ab263fd1a123af7a5 …pic.twitter.com/QdnICU4VAm

          2 replies 0 retweets 0 likes
        2. Bohumír Zámečník‏ @bzamecnik 30 Oct 2017
          Replying to @xpasky @fchollet @RossumAi

          Actually, the code is at https://github.com/rossumai/keras-multi-gpu/blob/master/keras_tf_multigpu/callbacks.py#L10 … + examples: https://github.com/rossumai/keras-multi-gpu/tree/master/keras_tf_multigpu/examples/bzamecnik/staging_area …

          0 replies 0 retweets 1 like
          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