Face Verification on LFW
9,980Mean AccuracyViT-8R
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=WebFace4M2026.06 | 9,980 | — | — | |
| ViTBackbone=ViT-B, Train data=WebFace4M2026.06 | 9,973 | — | — | |
| SwinFaceBackbone=Swin-T, Train data=MS1MV22026.06 | 99.87 | — | — | |
| HUMAN-Fusion2019.08 | 99.85 | — | — | |
| VarGFaceNetParams.(M)=5.0, MFLOPS=10222021.08 | 99.85 | — | — | |
| GroupFaceTraining Dataset=clean MS1M [7], [4]2021.09 | 99.85 | — | — | |
| AFRN2022.03 | 99.85 | — | — | |
| GroupFace2022.03 | 99.85 | — | — | |
| BroadFace2022.03 | 99.85 | — | — | |
| PFCTraining Dataset=WebFace4M, Backbone=ResNet-100, Sampling Ratio=0.32022.03 | 99.85 | — | — | |
| PFCTraining Dataset=WebFace42M, Backbone=ResNet-100, Sampling Ratio=0.32022.03 | 99.85 | — | — | |
| AFRN#Image=3.1M2018.01 | 99.85 | — | — | |
| GroupFace#Image=5.8M2018.01 | 99.85 | — | — | |
| BroadFace#Image=5.8M2018.01 | 99.85 | — | — | |
| VarGFaceNetMPARAMS=5.0, MFLOPS=10222023.07 | 99.85 | — | — | |
| VarGFaceNetCategory=Baseline, Params [M]=5.02023.07 | 99.85 | — | — | |
| GroupFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.85 | — | — | |
| BroadFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.85 | — | — | |
| GIFTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.85 | — | — | |
| GIFTrain Set=WebFace4M, Backbone=ResNet-1002025.05 | 99.85 | — | — | |
| BroadFaceTraining Dataset=MS1MV22026.04 | 99.85 | — | — | |
| FunFaceTraining Dataset=WebFace4M2026.04 | 99.85 | — | — | |
| BroadFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 99.85 | — | — | |
| TransFaceBackbone=TransFace-B, Train data=MS1MV22026.06 | 99.85 | — | — | |
| TransFace++-BBackbone=ViT-B, Train data=MS1MV22026.06 | 99.85 | — | — | |
| KP-RPE-8R*Backbone=ViT-B+KP-RPE + 8 Registers, Train data=WebFace4M2026.06 | 99.85 | — | — | |
| Prodpoly-ResNet50Model=Prodpoly-ResNet502020.06 | 99.833 | — | — | |
| ArcFaceTraining Data=5.8M2019.04 | 99.83 | — | — | |
| ArcfaceTraining epoch=16, Params=65M, MAdds=12.1G, Batch size=5122019.08 | 99.83 | — | — | |
| MagFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 99.83 | — | — | |
| MagFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.83 | — | — | |
| Partial-FC-ArcFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.83 | — | — | |
| Partial-FC-CosFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.83 | — | — | |
| ArcFace2022.03 | 99.83 | — | — | |
| DUL2022.03 | 99.83 | — | — | |
| VPL2022.03 | 99.83 | — | — | |
| Virtual FC2022.03 | 99.83 | — | — | |
| WebFace12M2022.03 | 99.83 | — | — | |
| WebFace42M2022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace4M, Backbone=ResNet-100, Sampling Ratio=0.042022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace12M, Backbone=ResNet-100, Sampling Ratio=0.0132022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace12M, Backbone=ResNet-100, Sampling Ratio=0.32022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace42M, Backbone=ResNet-100, Sampling Ratio=0.0082022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace42M, Backbone=ResNet-200, Sampling Ratio=0.32022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace42M, Backbone=ViT-B, Sampling Ratio=0.32022.03 | 99.83 | — | — | |
| PFCTraining Dataset=WebFace42M, Backbone=ViT-L, Sampling Ratio=0.32022.03 | 99.83 | — | — | |
| ResNet100param=65.2M, Bits=FP32, Size (MB)=261.222022.06 | 99.83 | — | — | |
| BroadFaceTraining Set=IBUG500K, Backbone=ResNet100, #Image=11.96M2018.01 | 99.83 | — | — | |
| ArcFaceTraining Set=MS1MV3, Backbone=ResNet100, #Image=5.1M2018.01 | 99.83 | — | — | |
| ArcFaceTraining Set=IBUG500K, Backbone=ResNet100, #Image=11.96M2018.01 | 99.83 | — | — | |
| ArcFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.83 | — | — | |
| DULTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.83 | — | — | |
| MagFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.83 | — | — | |
| ArcFaceTrain Set=WebFace4M, Backbone=ResNet-1002025.05 | 99.83 | — | — | |
| GIFTrain Set=WebFace4M, Backbone=ViT-base2025.05 | 99.83 | — | — | |
| ArcFace(m=0.50)Training Dataset=MS1MV22026.04 | 99.83 | — | — | |
| MagFaceTraining Dataset=MS1MV22026.04 | 99.83 | — | — | |
| PartialFC(ArcFace)Training Dataset=MS1MV22026.04 | 99.83 | — | — | |
| PartialFC(CosFace)Training Dataset=MS1MV22026.04 | 99.83 | — | — | |
| VPL-ArcFaceTraining Dataset=MS1MV32026.04 | 99.83 | — | — | |
| AdaFaceTraining Dataset=MS1MV32026.04 | 99.83 | — | — | |
| FunFaceTraining Dataset=MS1MV32026.04 | 99.83 | — | — | |
| ArcFaceTraining Dataset=WebFace4M2026.04 | 99.83 | — | — | |
| ArcFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 99.83 | — | — | |
| MagFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 99.83 | — | — | |
| VPL-ArcFaceBackbone=ResNet101, Train data=MS1MV32026.06 | 99.83 | — | — | |
| AdaFaceBackbone=ResNet101, Train data=MS1MV32026.06 | 99.83 | — | — | |
| Part-based FRBackbone=Part fViT-B, Train data=MS1MV32026.06 | 99.83 | — | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=MS1MV32026.06 | 99.83 | — | — | |
| ArcFaceBackbone=ResNet101, Train data=WebFace4M2026.06 | 99.83 | — | — | |
| KP-RPEBackbone=ViT-B+KP-RPE, Train data=WebFace4M2026.06 | 99.83 | — | — | |
| PFEfuse+matchTraining Data=4.4M2019.04 | 99.82 | — | — | |
| ArcFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.82 | — | — | |
| ElasticFace-CosTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.82 | — | — | |
| ElasticFace-Arc+Training Dataset=MS1MV2 [7], [4]2021.09 | 99.82 | — | — | |
| PFE#Image=4.4M2018.01 | 99.82 | — | — | |
| ElasticFace (MS1MV2)Augmentation=false, Identities=85 K, Total training images=5,800 K2023.08 | 99.82 | — | — | |
| ResNet-100 ArcFaceCategory=Baseline, Params [M]=65.22023.07 | 99.82 | — | — | |
| ElasticFace-Arc+Training Dataset=MS1MV22026.04 | 99.82 | — | — | |
| AdaFace(m=0.4)Training Dataset=MS1MV22026.04 | 99.82 | — | — | |
| FunFaceTraining Dataset=MS1MV22026.04 | 99.82 | — | — | |
| AdaFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 99.82 | — | — | |
| ViTBackbone=ViT-B, Train data=MS1MV22026.06 | 99.82 | — | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=MS1MV22026.06 | 99.82 | — | — | |
| ArcFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 99.81 | — | — | |
| CosFace2022.03 | 99.81 | — | — | |
| CosFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 99.81 | — | — | |
| CosFace(m=0.35)Training Dataset=MS1MV22026.04 | 99.81 | — | — | |
| CosFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 99.81 | — | — | |
| DW-SeesawFaceNet V1Training epoch=16, Params=4.1M, MAdds=514M, Batch size=1282019.08 | 99.8 | — | — | |
| DW-SeesawFaceNet V2Training epoch=16, Params=4.2M, MAdds=526M, Batch size=1282019.08 | 99.8 | — | — | |
| MV-Arc-Softmax*Backbone=ResNet1002020.04 | 99.8 | — | — | |
| CurricularFaceBackbone=ResNet1002020.04 | 99.8 | — | — | |
| UniformFaceTraining Dataset=clean MS1M [7], [4] + VGGFace2 [2]2021.09 | 99.8 | — | — | |
| CurricularFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.8 | — | — | |
| Dyn-arcFaceTraining Dataset=clean MS1M [7], [4]2021.09 | 99.8 | — | — | |
| ElasticFace-ArcTraining Dataset=MS1MV2 [7], [4]2021.09 | 99.8 | — | — | |
| ElasticFace-Cos+Training Dataset=MS1MV2 [7], [4]2021.09 | 99.8 | — | — | |
| MV-Softmax2022.03 | 99.8 | — | — | |
| CurricularFace2022.03 | 99.8 | — | — |