Quantization in Layer's Input is Matter

Daning Cheng; WenGuang Chen

Quantization in Layer's Input is Matter

Machine Learning 2022-02-11 v1

Authors: Daning Cheng , WenGuang Chen

Abstract

In this paper, we will show that the quantization in layer's input is more important than parameters' quantization for loss function. And the algorithm which is based on the layer's input quantization error is better than hessian-based mixed precision layout algorithm.

Cite

@article{arxiv.2202.05137,
  title  = {Quantization in Layer's Input is Matter},
  author = {Daning Cheng and WenGuang Chen},
  journal= {arXiv preprint arXiv:2202.05137},
  year   = {2022}
}

Related papers

View all related →