Weiyang Liu

All the models we provided is under MIT license. For the CASIA dataset, I am not very sure about its license. But you can refer to [this](http://www.cbsr.ia.ac.cn/english/CASIA-WebFace/CASIA-WebFace_Agreements.pdf).

Thank you very much for your contributions. We will check the code ASAP.

I should be okay with the old official caffe, but I have not used the newest version. So I can not comment on that. I think with minor modifications, it...

SINGLE, DOUBLE, etc correspond to the case of m=1, m=2, etc. It is simply for the convenience of implementation.

It is a standard data processing procedure. I think it is used not only in deep neural nets. Imagine a simple case where you just classify two points in 2D...

@johnnysclai I don’t think it really matters. As long as your testing and training are consistent, you will be fine.

Not sure about which dataset you use. If you are using different dataset other than CASIA, the loss will likely be different. You need to adjust the hyperparameters.

It is exactly the uncertain factor you mentioned. There may be some problems in your downloaded cropped dataset, and there may be some mis-matches between the cropped dataset and the...

The detailed setting is available in the [training log](https://drive.google.com/file/d/0B_geeR2lTMegcWkxdVV4X1FOaFU/view) we released. The provided models are trained using the exact same setting in the repository. As for some other unpredictable issues...

You just need to use the pretrained model to extract the features for your two custom faces, then compute their cosine similarity, and finally use a threshold to determine whether...