Abstract
Nowadays, with the advancement of deep neural networks (DNNs) and the availability of large-scale datasets, the face recognition (FR) model has achieved exceptional performance. However, since the parameter magnitude of the fully connected (FC) layer directly depends on the number of identities in the dataset. If training the FR model on large-scale datasets, the size of the model parameter will be excessively huge, leading to substantial demand for computational resources, such as time and memory. This paper proposes the attention fully connected (AttFC) layer, which could significantly reduce computational resources. AttFC employs an attention loader to generate the generative class center (GCC), and dynamically store the class center with Dynamic Class Container (DCC). DCC only stores a small subset of all class centers in FC, thus its parameter count is substantially less than the FC layer. Also, training face recognition models on large-scale datasets with one GPU often encounter out-of-memory (OOM) issues. AttFC overcomes this and achieves comparable performance to state-of-the-art methods.
Abstract (translated)
如今，随着深度神经网络（DNN）的发展和大规模数据集的可用性，人脸识别（FR）模型已经取得了卓越的成绩。然而，由于全连接层（FC）参数的数量直接依赖于数据集中身份的数量，当在大规模数据集上训练FR模型时，模型参数的数量会变得极其庞大，从而对计算资源的需求大幅增加，如时间和内存等。本文提出了注意全连接（AttFC）层，该方法可以显著减少计算资源需求。AttFC使用注意力加载器来生成生成类中心（GCC），并利用动态类别容器（DCC）动态存储类中心。DCC仅存储全连接层中所有类中心的一个小子集，因此其参数数量远小于传统的全连接层。此外，在具有大规模数据集的情况下，用单个GPU训练人脸识别模型常常会遇到内存不足的问题。AttFC解决了这一问题，并且在性能上与最先进的方法相当。
URL
https://arxiv.org/abs/2503.06839
