by Feng Wang, Weiyang Liu, Haijun Liu, Jian Cheng
The paper is available as a technical report at arXiv.
In this work, we design a new loss function which merges the merits of both NormFace and SphereFace. It is much easier to understand and train, and outperforms the previous state-of-the-art loss function (SphereFace) by 2-5% on MegaFace.
If you find AM-Softmax useful in your research, please consider to cite:
@article{Wang_2018_amsoftmax,
title = {Additive Margin Softmax for Face Verification},
author = {Feng Wang, Weiyang Liu, Haijun Liu, Jian Cheng},
journal = {arXiv preprint arXiv:1801.05599},
year = {2018}
}
Requirements: My Caffe version https://github.com/happynear/caffe-windows. This version can also be compiled in Linux.
The prototxt file is in ./prototxt. The batch size is set to 256. If your GPU's memory is not sufficient enough, you may set iter_size: 2 in face_solver.prototxt and batch_size: 128 in face_train_test.prototxt.
The dataset used for training is CASIA-Webface. We removed 59 identities that are duplicated with LFW (17) and MegaFace Set 1 (42). This is why the final inner-product layer's output is 10516. The list of the duplicated identities can be found in https://github.com/happynear/FaceDatasets.
All other settings are the same with SphereFace. Please refer to the details in SphereFace's repository.
Feature normalized, s=30, m=0.35: OneDrive, Baidu Yun .
See our arXiv technical report.