@article{squeezingsgdparallelizationperformancein,
  title = {Squeezing SGD Parallelization Performance in Distributed Training Using Delayed Averaging},
  author = {Pengcheng Li and Yixin Guo and Yawen Zhang and Qinggang Zhou},
  year = {2021},
  url = {https://openreview.net/forum?id=DtfrnB1fiX},
}