use caffe to training a deepid2 net with softmax and contrastive loss, adding identity2verify_layer to translate a single sample into pairwise, thus mix the softmax and contrastive layer together
guowenzhe / deepid2 Goto Github PK
View Code? Open in Web Editor NEWThis project forked from wjgaas/deepid2
use caffe to training a deepid2 net with softmax and contrastive loss, adding identity2verify_layer to translate a single sample into pairwise, thus mix the softmax and contrastive layer together