Open chensteven opened 6 years ago
There are links to the network architecture and training code here: https://github.com/ageitgey/face_recognition/wiki/Face-Recognition-Accuracy-Problems#question-can-i-re-train-the-face-encoding-model-to-make-it-more-accurate-for-my-images
pytorch main point is TripletLoss
128-D embedding, embedding size = 128
https://github.com/tbmoon/facenet
Is there a training code or example available for the model dlib_face_recognition_resnet_model_v1 from https://github.com/davisking/dlib-models?
Also curious if there are any document on what the ResNet architecture look like and etc.
Thanks!