Distributed TensorFlow (TensorFlow Dev Summit 2018)

แชร์
ฝัง
  • เผยแพร่เมื่อ 18 พ.ย. 2024

ความคิดเห็น • 13

  • @DavidEllisonPhD
    @DavidEllisonPhD 6 ปีที่แล้ว +2

    Thanks for including links to articles in your talk. I was especially happy to learn about your future plans with distributed TensorFlow and Horovod. Thank you for mentioning that.

  • @aleksandrtsyrulnev7021
    @aleksandrtsyrulnev7021 6 ปีที่แล้ว +1

    the best software engineer in the world

  • @jackshi9500
    @jackshi9500 6 ปีที่แล้ว

    This is awesome!
    Good to hear that it is not tied to estimators. It is very hacky to build reinforcement learning systems with Estimator API since "input" is often generated and depend on previous model output, label could be extracted from internal state (prioritized memory for DQN), and number of steps depend on environment feedback.
    Looking forward for this to become available for Graph Session API !

  • @californiaesnuestra
    @californiaesnuestra 6 ปีที่แล้ว

    Great presentation! Interesting to see a mention of Uber Horovod as a way to embrace the Open Source commitment from Google/TF team.

  • @ashokvasthav8906
    @ashokvasthav8906 4 ปีที่แล้ว

    I wabted to see on example for mirrored strategy for say imagenet. R there any links

  • @captvalerii
    @captvalerii 6 ปีที่แล้ว +1

    Steady so !

  • @jacojoubert12
    @jacojoubert12 5 ปีที่แล้ว

    Does someone have information/links on how the inter-node communication works? MPI? NCCL? Both? Were these results obtained using TCP over LAN or Infiniband/RoCE?

  • @clydexu6129
    @clydexu6129 5 ปีที่แล้ว

    honestly i have little idea how the distributed gradients can be combined together.

    • @RobertLugg
      @RobertLugg 4 ปีที่แล้ว

      You might imaging a few ways. But don't think there is anything super high-tech. You need to "resolve" the gradient values. This could just mean taking the average, the mins, the maxes. Or, you might try to see which give the best cost and then use those gradients for all nodes.

  • @deeplearningpartnership
    @deeplearningpartnership 6 ปีที่แล้ว +4

    Great talk.

  • @iradukundajeandedieu4750
    @iradukundajeandedieu4750 6 ปีที่แล้ว

    Great Talk

  • @yilinliu4948
    @yilinliu4948 6 ปีที่แล้ว

    COOL, THANKS