본문 바로가기

쓰기

 
발표자 박은미 
발표일자 2021-04-21 
저자 Andrew Brock 
학회명  
논문지  

제목: High-Performance Large-Scale Image Recognition Without Normalization

저자: Andrew Brock, Soham De, Samuel L. Smith, Karen Simonyan 


20210421_165651.png

time.png

For completeness, in Table 6 of the Appendix we also report the performance of our model architectures when trained with batch normalization instead of the NF strategy. These models achieve slightly lower test accuracies than their NF counterparts and they are between 20% and 40% slower to train, even when using highly optimized batch normalization implementations without cross-replica syncing.

Drop here!
Drop here!
Drop here!
Drop here!
Drop here!

    2022

      GPipe: Easy Scaling with Micro-Batch Pipeline Parallelism
      2022.02.22
      발표자: 조영성     발표일자: 2022-02-22     저자: Yanping Huang, Youlong Cheng, Ankur Bapna, Orhan Firat, Mia Xu Chen, Dehao Chen, HyoukJoong Lee, Jiquan Ngiam, Quoc V. Le, Yonghui Wu, Zhifeng Chen     논문지: https://arxiv.org/abs/1811.06965    
      Contrastive Code Representation Learning
      2022.02.15
      발표자: 최윤석     발표일자: 2022-02-15     저자: Paras Jain, Ajay Jain, Tianjun Zhang, Pieter Abbeel, Joseph E. Gonzalez, Ion Stoica     학회명: EMNLP 2021