English

A Robust Initialization of Residual Blocks for Effective ResNet Training without Batch Normalization

Machine Learning 2023-11-07 v2

Abstract

Batch Normalization is an essential component of all state-of-the-art neural networks architectures. However, since it introduces many practical issues, much recent research has been devoted to designing normalization-free architectures. In this paper, we show that weights initialization is key to train ResNet-like normalization-free networks. In particular, we propose a slight modification to the summation operation of a block output to the skip-connection branch, so that the whole network is correctly initialized. We show that this modified architecture achieves competitive results on CIFAR-10, CIFAR-100 and ImageNet without further regularization nor algorithmic modifications.

Cite

@article{arxiv.2112.12299,
  title  = {A Robust Initialization of Residual Blocks for Effective ResNet Training without Batch Normalization},
  author = {Enrico Civitelli and Alessio Sortino and Matteo Lapucci and Francesco Bagattini and Giulio Galvan},
  journal= {arXiv preprint arXiv:2112.12299},
  year   = {2023}
}

Comments

16 pages (4 pages of supplementary material), 9 figures, 2 table

R2 v1 2026-06-24T08:28:56.262Z