Face Verification on CALFW
9,592AccuracyViT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ViTBackbone=ViT-B, Train data=WebFace4M2026.06 | 9,592 | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=WebFace4M2026.06 | 9,592 | — | |
| GIFTrain Set=WebFace4M, Backbone=ViT-base2025.05 | 97.19 | — | |
| GIFTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 96.94 | — | |
| GIFTrain Set=WebFace4M, Backbone=ResNet-1002025.05 | 96.85 | — | |
| TopoFRTraining Data=MS1MV2, Backbone=ResNet-100, Venue=NeurIPS242024.10 | 96.44 | — | |
| TopoFR+Training Data=MS1MV2, Backbone=ResNet-100, Venue=NeurIPS242024.10 | 96.42 | — | |
| TopoFRTraining Data=MS1MV2, Backbone=ResNet-50, Venue=NeurIPS242024.10 | 96.25 | — | |
| FunFaceTraining Dataset=MS1MV22026.04 | 96.25 | — | |
| TopoFR+Training Data=MS1MV2, Backbone=ResNet-50, Venue=NeurIPS242024.10 | 96.24 | — | |
| Prodpoly-ResNet50Model=Prodpoly-ResNet502020.06 | 96.233 | — | |
| CurricularFaceBackbone=ResNet1002020.04 | 96.2 | — | |
| GroupFaceTraining Dataset=clean MS1M [7], [4]2021.09 | 96.2 | — | |
| CurricularFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 96.2 | — | |
| Partial-FC-CosFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 96.2 | — | |
| BroadFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=ECCV202024.10 | 96.2 | — | |
| CurricularFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR202024.10 | 96.2 | — | |
| GroupFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 96.2 | — | |
| CurricularFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 96.2 | — | |
| BroadFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 96.2 | — | |
| CurricularFaceTraining Dataset=MS1MV22026.04 | 96.2 | — | |
| BroadFaceTraining Dataset=MS1MV22026.04 | 96.2 | — | |
| PartialFC(CosFace)Training Dataset=MS1MV22026.04 | 96.2 | — | |
| BroadFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 96.2 | — | |
| CurricularFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 96.2 | — | |
| TransFaceBackbone=TransFace-B, Train data=MS1MV22026.06 | 96.2 | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=MS1MV22026.06 | 96.2 | — | |
| CosFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 96.18 | — | |
| Partial-FC-ArcFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 96.18 | — | |
| ElasticFace-Cos+Training Dataset=MS1MV2 [7], [4]2021.09 | 96.18 | — | |
| ResNet-100 ElasticFace (Cos+)Category=Baseline, Params [M]=65.22023.07 | 96.18 | — | |
| ElasticFace-Cos+Training Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR222024.10 | 96.18 | — | |
| IIC-AdaFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=ICLR242024.10 | 96.18 | — | |
| ElasticFace-Cos+Training Dataset=MS1MV22026.04 | 96.18 | — | |
| PartialFC(ArcFace)Training Dataset=MS1MV22026.04 | 96.18 | — | |
| ElasticFace-Cos+Backbone=ResNet101, Train data=MS1MV22026.06 | 96.18 | — | |
| Face TransformerBackbone=ViT-P12S8, Train data=MS1MV22026.06 | 96.18 | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=MS1MV32026.06 | 96.18 | — | |
| ElasticFace-ArcTraining Dataset=MS1MV2 [7], [4]2021.09 | 96.17 | — | |
| ElasticFace-Arc+Training Dataset=MS1MV2 [7], [4]2021.09 | 96.17 | — | |
| ElasticFace-Arc+Training Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR222024.10 | 96.17 | — | |
| ElasticFace-Arc+Training Dataset=MS1MV22026.04 | 96.17 | — | |
| MagFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 96.15 | — | |
| MagFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 96.15 | — | |
| MagFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR212024.10 | 96.15 | — | |
| MagFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 96.15 | — | |
| MagFaceTraining Dataset=MS1MV22026.04 | 96.15 | — | |
| FunFaceTraining Dataset=MS1MV32026.04 | 96.15 | — | |
| FunFaceTraining Dataset=WebFace4M2026.04 | 96.15 | — | |
| MagFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 96.15 | — | |
| ResNet100Bits=FP322022.06 | 96.13 | — | |
| SCF-ArcFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR212024.10 | 96.12 | — | |
| VPL-ArcFaceTraining Dataset=MS1MV32026.04 | 96.12 | — | |
| VPL-ArcFaceBackbone=ResNet101, Train data=MS1MV32026.06 | 96.12 | — | |
| ViTBackbone=ViT-B, Train data=MS1MV32026.06 | 96.12 | — | |
| MV-Arc-Softmax*Backbone=ResNet1002020.04 | 96.1 | — | |
| ResNet50Bits=FP322022.06 | 96.1 | — | |
| MV-Softmax2018.01 | 96.1 | — | |
| ArcFaceBackbone=R100, Training Dataset=IBUG500K2018.01 | 96.1 | — | |
| MV-SoftmaxTraining Data=MS1MV2, Backbone=ResNet-100, Venue=AAAI202024.10 | 96.1 | — | |
| MV-SoftMaxTraining Dataset=MS1MV22026.04 | 96.1 | — | |
| SwinFaceBackbone=Swin-T, Train data=MS1MV22026.06 | 96.1 | — | |
| AdaFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR222024.10 | 96.08 | — | |
| AdaFace(m=0.4)Training Dataset=MS1MV22026.04 | 96.08 | — | |
| ViT-FREE11Exit Depth=11, Fine-tuning Protocol=None, GFLOPs=10.552026.06 | 96.08 | — | |
| AdaFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 96.08 | — | |
| DW-SeesawFaceNet V1Training epoch=16, Params=4.1M, MAdds=514M, Batch size=1282019.08 | 96.07 | — | |
| AdaFaceTraining Data=MS1MV2, Backbone=ResNet-50, Venue=CVPR222024.10 | 96.07 | — | |
| ResNet100Quantization data=Real data, Bits=w8a82022.06 | 96.05 | — | |
| CurricularFace2018.01 | 96.05 | — | |
| AdaFaceTraining Dataset=WebFace4M2026.04 | 96.05 | — | |
| ViT-FREE12Exit Depth=12, Fine-tuning Protocol=None, GFLOPs=11.502026.06 | 96.05 | — | |
| AdaFaceBackbone=ResNet101, Train data=WebFace4M2026.06 | 96.05 | — | |
| ElasticFace-CosTraining Dataset=MS1MV2 [7], [4]2021.09 | 96.03 | — | |
| AdaFaceTrain Set=WebFace4M, Backbone=ViT-base2025.05 | 96.03 | — | |
| QuantFaceBackbone=ResNet100, Quantization data=Synthetic data, Bits=w8a82022.06 | 96.02 | — | |
| ArcFaceBackbone=R100, Training Dataset=MS1MV32018.01 | 96.02 | — | |
| AdaFaceTraining Dataset=MS1MV32026.04 | 96.02 | — | |
| AdaFaceBackbone=ResNet101, Train data=MS1MV32026.06 | 96.02 | — | |
| ResNet50Quantization data=Real data, Bits=w8a82022.06 | 96 | — | |
| ArcFaceTrain Set=WebFace4M, Backbone=ResNet-1002025.05 | 96 | — | |
| ArcFaceTraining Dataset=WebFace4M2026.04 | 96 | — | |
| ViTBackbone=ViT-B, Train data=MS1MV22026.06 | 96 | — | |
| ArcFaceBackbone=ResNet101, Train data=WebFace4M2026.06 | 96 | — | |
| KP-RPEBackbone=ViT-B+KP-RPE, Train data=WebFace4M2026.06 | 96 | — | |
| KP-RPE-8R*Backbone=ViT-B+KP-RPE + 8 Registers, Train data=WebFace4M2026.06 | 96 | — | |
| DW-SeesawFaceNet V2Training epoch=16, Params=4.2M, MAdds=526M, Batch size=1282019.08 | 95.98 | — | |
| ArcFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 95.96 | — | |
| ResNet50Model=ResNet502020.06 | 95.917 | — | |
| MagFaceTraining Data=MS1MV2, Backbone=ResNet-50, Venue=CVPR212024.10 | 95.88 | — | |
| QuantFaceBackbone=ResNet50, Quantization data=Synthetic data, Bits=w8a82022.06 | 95.87 | — | |
| ResNet50Quantization data=Real data, Bits=w6a62022.06 | 95.87 | — | |
| KP-RPE*Backbone=ViT-B+KP-RPE, Train data=WebFace4M2026.06 | 95.83 | — | |
| GhostFaceNetV1-1 KUCategory=2-5 MP, Params [M]=4.092023.07 | 95.77 | 1 | |
| CosFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 95.76 | — | |
| CosFace(m=0.35)Training Dataset=MS1MV22026.04 | 95.76 | — | |
| CosFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 95.76 | — | |
| upper-boundarysource=reproduced by authors2021.05 | 95.75 | — | |
| CosFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR182024.10 | 95.75 | — | |
| ResNet18Quantization data=Real data, Bits=w8a82022.06 | 95.72 | — |