sphereface的pytorch实现代码，2017的一篇cvpr，SphereFace: Deep Hypersphere Embedding for Face Recognition，继centerloss之后又一大作。 文章主要提出了归一化权值（normalize weights and zero biases）和角度间距（angular margin），基于这2个点，对传统的softmax进行了改进，从而实现了，最大类内距离小于最小类间距离的识别标准。
Convolutional neural networks (CNNs) have been widely used in computer vision community, significantly improving the state-ofthe- art. In most of the available CNNs, the softmax loss function is used as the supervision signal to train the deep model
