30.09.2019 · Training a face Recognizer using ResNet50 + ArcFace in TensorFlow 2.0 The aim of this project is to train a state of art face recognizer using TensorFlow 2.0. The architecture chosen is a modified version of ResNet50 and the loss function used is ArcFace, both originally developed by deepinsight in mxnet.
Verification performance (%) of ArcFace using ResNet- 101 [44] trained on VGGFace2 [42] and VGGFace2 8631 Races with syntesised images of non-Caucasian subjects ...
arcface-insightface : Arcface model (Resnet100 backbone) from Insightface. resnet50-msceleb-arcface-2021 : Resnet Arcface model trained with MSCeleb dataset ...
arcface-pytorch / models / resnet.py / Jump to. Code definitions. conv3x3 Function BasicBlock Class __init__ Function forward Function IRBlock Class __init__ Function ...
ResNet-100 is used for training. ... As reported in Table4, sub-center ArcFace trained on noisy MS1MV0 achieves comparable performance compared to ArcFace ...
arcface-pytorch / models / resnet.py / Jump to. Code definitions. conv3x3 Function BasicBlock Class __init__ Function forward Function IRBlock Class __init__ Function forward Function Bottleneck Class __init__ Function forward Function SEBlock Class __init__ Function forward Function ResNetFace Class __init__ Function _make_layer Function ...
Sep 30, 2019 · The architecture chosen is a modified version of ResNet50 and the loss function used is ArcFace, both originally developed by deepinsight in mxnet. The dataset used for training is the CASIA-Webface MS1M-ArcFace dataset used in insightface, and it is available their dataset zoo. The images are aligned using mtcnn and cropped to 112x112.
In this paper, we propose an Additive Angular Margin Loss (ArcFace) to obtain highly discriminative features for face recognition. The proposed ArcFace has ...