qidiso / mobilefacenet-mxnet
基于insightface训练mobilefacenet的调试步骤,更改模型后一层训练结果为99.683% in lfw and 96.717 in agedb. Now pls move to the new mobilefacenet-V2(http://github.com/qidiso/mobilefacenet-V2)
☆11Updated 6 years ago
Alternatives and similar repositories for mobilefacenet-mxnet:
Users that are interested in mobilefacenet-mxnet are comparing it to the libraries listed below
- a simplify and accuracy-maintain model of ResNet-50 by the Invert Residual Construction☆37Updated 6 years ago
- pycaffe version of RSA 'Recurrent Scale Approximation for Object Detection in CNN'☆32Updated 7 years ago
- Diagonalwise Refactorization: An Efficient Training Method for Depthwise Convolutions (in Caffe)☆34Updated 7 years ago
- My implementation of <Range Loss for Deep Face Recognition with Long-tail>☆22Updated 7 years ago
- ☆16Updated 6 years ago
- reproduce faceboxes☆53Updated 6 years ago
- The network of the faceboxes☆15Updated 7 years ago
- Facial landmarks training by mxnet(gluon). Use for face alignment and so on.☆40Updated 7 years ago
- A simplify version of mobilenet, with less group and feature maps, trained on Imagenet.