AlfredXiangWu
IJB-A test protocol and result
Dear @AlfredXiangWu , I found IJB-A test results on your paper, but I can't find it in your code. Can you help me find it or guide me to reproduce the results on IJB-A?
Thank you so much!
Refer to https://github.com/deepinsight/insightface/tree/master/recognition/evaluation/ijb
Refer to https://github.com/deepinsight/insightface/tree/master/recognition/evaluation/ijb
Hi Alfred XiangWu, Thank you for your reply
I read this reference, but I think it was used for IJB-B and IJB-C. Do you have any reference or your demo on IJB-A evaluation? Besides, I also want to ask if Caffe Model A in ''AlfredXiangWu/face_verification_experiment'' is a pre-trained model with LightCNN-4 layers, right?
Unfortunately, I could not find my evaluation code of IJB-A. As far as I know, IJB-A evaluation is similar to IJB-B and IJB-C. You can refer to the imeplement of insightface to test IJB-A.
The caffe model A is LightCNN-4 layer model, which is trained on the CASIA-WebFace dataset.
Do you have Pytorch version of this model (Model A), I can not find it on this repo.