A Robust Initialization of Residual Blocks for Effective ResNet Training without Batch Normalization
This addresses the practical issues of Batch Normalization for researchers and practitioners in deep learning, though it is incremental as it builds on existing normalization-free architectures.
The paper tackled the problem of training ResNet-like networks without Batch Normalization by proposing a modified initialization for residual blocks, achieving competitive results on CIFAR-10, CIFAR-100, and ImageNet datasets.
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.