Residual Attention Networks for Image Classification

This work combines components of the Inception architecture with ideas from ResNet, and from Visual Attention Networks, to achieve reasonable performance on the given dataset, known as Tiny ImageNet.