Convolutional neural networks have significantly boosted the performance of face recognition in recent years due to its high capacity in learning discriminative features. To enhance the discriminative power of the Softmax loss, multiplicative angular margin and additive cosine margin incorporate angular margin and cosine margin into the loss functions, respectively. In this paper, we propose a novel supervisor signal, additive angular margin (ArcFace), which has a better geometrical interpretation than supervision signals proposed so far. Specifically, the proposed ArcFace $\cos(\theta + m)$ directly maximise decision boundary in angular (arc) space based on the L2 normalised weights and features. Compared to multiplicative angular margin $\cos(m\theta)$ and additive cosine margin $\cos\theta-m$, ArcFace can obtain more discriminative deep features. We also emphasise the importance of network settings and data refinement in the problem of deep face recognition. Extensive experiments on several relevant face recognition benchmarks, LFW, CFP and AgeDB, prove the effectiveness of the proposed ArcFace. Most importantly, we get state-of-art performance in the MegaFace Challenge in a totally reproducible way. We make data, models and training/test code public available~\footnote{https://github.com/deepinsight/insightface}.