EfficientNet

  • EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks, Mingxing Tan, Quoc V.Le 2019
  • 다양한 Convolution Net들이 출현중인데 한가지의 공통점이 존재. 깊이를 깊게 하거나 넓이를 넓게 하거나 구조를 다양하게 해서 좀더 정확도가 높은 지점을 찾는다는것
  • EfficientNet은 이 구조의 효율성을 찾아내 정확하게 예측력과 컴퓨팅파워를 절묘하게 균형을 맞춰서 SOTA를 달성
  • 구조가 복잡할수록 비례적으로 정확도는 증가하나 복잡도도 또한 증가.
  • B0(빠름), B4(효율좋음), B7(예측력정확) 정도가 주로 쓰임
부분에서 효율적인 구조를 찾았기 때문에, 앞으로는 EfficientNet을 빼놓고 벤치를 논하기가 불가능할듯 하다.

10 thoughts on “EfficientNet

  1. I’m not positive where you are getting your information, but good topic.

    I must spend some time finding out more or figuring
    out more. Thanks for magnificent info I used to be on the lookout for this info for my
    mission.

    Here is my web site: g give

  2. When some one searches for his essential thing, so he/she
    wishes to be available that in detail, so that thing is maintained over here.

    Feel free to surf to my web blog :: g look

  3. Wow, fantastic weblog layout! How long have you ever been blogging for?
    you made blogging look easy. The whole glance of your web
    site is excellent, as neatly as the content!

    my page – cbd oil (tinyurl.com)

  4. Undeniably believe that which you stated. Your
    favorite reason seemed to be on the net the easiest
    thing to be aware of. I say to you, I certainly get annoyed while people
    consider worries that they plainly don’t know about.
    You managed to hit the nail upon the top and defined out the whole thing
    without having side-effects , people can take a signal. Will likely
    be back to get more. Thanks

    My blog post cbd oil [ww2.telechat.info]

댓글 남기기