Face Verification on CPLFW
9,497AccuracyViT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ViTBackbone=ViT-B, Train data=WebFace4M2026.06 | 9,497 | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=WebFace4M2026.06 | 9,493 | — | |
| GIFTrain Set=WebFace4M, Backbone=ViT-base2025.05 | 95.97 | — | |
| KP-RPEBackbone=ViT-B+KP-RPE, Train data=WebFace4M2026.06 | 95.4 | — | |
| KP-RPE*Backbone=ViT-B+KP-RPE, Train data=WebFace4M2026.06 | 95.16 | — | |
| KP-RPE-8R*Backbone=ViT-B+KP-RPE + 8 Registers, Train data=WebFace4M2026.06 | 95.08 | — | |
| GIFTrain Set=WebFace4M, Backbone=ResNet-1002025.05 | 95.03 | — | |
| AdaFaceTrain Set=WebFace4M, Backbone=ViT-base2025.05 | 94.97 | — | |
| FC-R100Training Dataset=Webface42M, Backbone=ResNet-1002021.05 | 94.9 | — | |
| F2C-R100Training Dataset=Webface42M, Backbone=ResNet-1002021.05 | 94.85 | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=MS1MV32026.06 | 94.68 | — | |
| ViT-8RBackbone=ViT-B + 8 Registers, Train data=MS1MV22026.06 | 94.65 | — | |
| AdaFaceTraining Dataset=WebFace4M2026.04 | 94.63 | — | |
| FunFaceTraining Dataset=WebFace4M2026.04 | 94.63 | — | |
| ViTBackbone=ViT-B, Train data=MS1MV32026.06 | 94.63 | — | |
| AdaFaceBackbone=ResNet101, Train data=WebFace4M2026.06 | 94.63 | — | |
| GIFTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 94.45 | — | |
| ArcFaceTrain Set=WebFace4M, Backbone=ResNet-1002025.05 | 94.35 | — | |
| ArcFaceTraining Dataset=WebFace4M2026.04 | 94.35 | — | |
| ArcFaceBackbone=ResNet101, Train data=WebFace4M2026.06 | 94.35 | — | |
| ViTBackbone=ViT-B, Train data=MS1MV22026.06 | 94.27 | — | |
| FunFaceTraining Dataset=MS1MV32026.04 | 94.12 | — | |
| ViT-FREE11Exit Depth=11, Fine-tuning Protocol=None, GFLOPs=10.552026.06 | 94.1 | — | |
| ViT-FREE12Exit Depth=12, Fine-tuning Protocol=None, GFLOPs=11.502026.06 | 94 | — | |
| AdaFaceTraining Dataset=MS1MV32026.04 | 93.93 | — | |
| AdaFaceBackbone=ResNet101, Train data=MS1MV32026.06 | 93.93 | — | |
| FunFaceTraining Dataset=MS1MV22026.04 | 93.83 | — | |
| TopoFRTraining Data=MS1MV2, Backbone=ResNet-100, Venue=NeurIPS242024.10 | 93.79 | — | |
| TopoFR+Training Data=MS1MV2, Backbone=ResNet-100, Venue=NeurIPS242024.10 | 93.78 | — | |
| AdaFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR222024.10 | 93.53 | — | |
| AdaFace(m=0.4)Training Dataset=MS1MV22026.04 | 93.53 | — | |
| AdaFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 93.53 | — | |
| FC-R50Training Dataset=Glint360k, Backbone=ResNet-502021.05 | 93.48 | — | |
| IIC-AdaFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=ICLR242024.10 | 93.48 | — | |
| VPL-ArcFaceTraining Dataset=MS1MV32026.04 | 93.45 | — | |
| VPL-ArcFaceBackbone=ResNet101, Train data=MS1MV32026.06 | 93.45 | — | |
| ArcFaceBackbone=R100, Training Dataset=IBUG500K2018.01 | 93.43 | — | |
| SwinFaceBackbone=Swin-T, Train data=MS1MV22026.06 | 93.42 | — | |
| TopoFRTraining Data=MS1MV2, Backbone=ResNet-50, Venue=NeurIPS242024.10 | 93.38 | — | |
| TopoFR+Training Data=MS1MV2, Backbone=ResNet-50, Venue=NeurIPS242024.10 | 93.36 | — | |
| Prodpoly-ResNet50Model=Prodpoly-ResNet502020.06 | 93.317 | — | |
| ElasticFace-Arc+Training Dataset=MS1MV2 [7], [4]2021.09 | 93.28 | — | |
| ElasticFace-Arc+Training Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR222024.10 | 93.28 | — | |
| ElasticFace-Arc+Training Dataset=MS1MV22026.04 | 93.28 | — | |
| ElasticFace-ArcTraining Dataset=MS1MV2 [7], [4]2021.09 | 93.27 | — | |
| ElasticFace-Cos+Training Dataset=MS1MV2 [7], [4]2021.09 | 93.23 | — | |
| ResNet-100 ElasticFace (Cos+)Category=Baseline, Params [M]=65.22023.07 | 93.23 | — | |
| ElasticFace-Cos+Training Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR222024.10 | 93.23 | — | |
| ElasticFace-Cos+Training Dataset=MS1MV22026.04 | 93.23 | — | |
| ElasticFace-Cos+Backbone=ResNet101, Train data=MS1MV22026.06 | 93.23 | — | |
| ResNet100Bits=FP322022.06 | 93.22 | — | |
| ArcFaceBackbone=R100, Training Dataset=MS1MV32018.01 | 93.21 | — | |
| GroupFaceBackbone=ResNet101, Training Dataset=MS1MV22021.05 | 93.17 | — | |
| GroupFaceTraining Dataset=clean MS1M [7], [4]2021.09 | 93.17 | — | |
| ElasticFace-CosTraining Dataset=MS1MV2 [7], [4]2021.09 | 93.17 | — | |
| BroadFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=ECCV202024.10 | 93.17 | — | |
| GroupFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 93.17 | — | |
| BroadFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 93.17 | — | |
| BroadFaceTraining Dataset=MS1MV22026.04 | 93.17 | — | |
| BroadFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 93.17 | — | |
| SCF-ArcFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR212024.10 | 93.16 | — | |
| CosFace (ours)Backbone=ResNet101, Training Dataset=MS1MV22021.05 | 93.15 | — | |
| CurricularFaceBackbone=ResNet1002020.04 | 93.13 | — | |
| CurricularFaceBackbone=ResNet101, Training Dataset=MS1MV22021.05 | 93.13 | — | |
| CurricularFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 93.13 | — | |
| CurricularFace2018.01 | 93.13 | — | |
| CurricularFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR202024.10 | 93.13 | — | |
| CurricularFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 93.13 | — | |
| CurricularFaceTraining Dataset=MS1MV22026.04 | 93.13 | — | |
| CurricularFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 93.13 | — | |
| PartialFC-r1.0Backbone=ResNet101, Training Dataset=MS1MV2, Sampling Rate=1.02021.05 | 93.1 | — | |
| Partial-FC-CosFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 93.1 | — | |
| PartialFC(CosFace)Training Dataset=MS1MV22026.04 | 93.1 | — | |
| Face TransformerBackbone=ViT-P12S8, Train data=MS1MV22026.06 | 93.08 | — | |
| Partial-FC-ArcFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 93 | — | |
| PartialFC(ArcFace)Training Dataset=MS1MV22026.04 | 93 | — | |
| ResNet100Quantization data=Real data, Bits=w8a82022.06 | 92.92 | — | |
| TransFaceBackbone=TransFace-B, Train data=MS1MV22026.06 | 92.92 | — | |
| PartialFC-r0.1Backbone=ResNet101, Training Dataset=MS1MV2, Sampling Rate=0.12021.05 | 92.9 | — | |
| QuantFaceBackbone=ResNet100, Quantization data=Synthetic data, Bits=w8a82022.06 | 92.9 | — | |
| DCQBackbone=ResNet101, Training Dataset=MS1MV22021.05 | 92.87 | — | |
| MagFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 92.87 | — | |
| MagFaceTraining Dataset=MS1MV2 [7], [4]2021.09 | 92.87 | — | |
| MagFaceTraining Data=MS1MV2, Backbone=ResNet-100, Venue=CVPR212024.10 | 92.87 | — | |
| MagFaceTrain Set=MS1MV2, Backbone=ResNet-1002025.05 | 92.87 | — | |
| MagFaceTraining Dataset=MS1MV22026.04 | 92.87 | — | |
| MagFaceBackbone=ResNet101, Train data=MS1MV22026.06 | 92.87 | — | |
| MV-Arc-Softmax*Backbone=ResNet1002020.04 | 92.83 | — | |
| MV-Softmax2018.01 | 92.83 | — | |
| AdaFaceTraining Data=MS1MV2, Backbone=ResNet-50, Venue=CVPR222024.10 | 92.83 | — | |
| MV-SoftmaxTraining Data=MS1MV2, Backbone=ResNet-100, Venue=AAAI202024.10 | 92.83 | — | |
| MV-SoftMaxTraining Dataset=MS1MV22026.04 | 92.83 | — | |
| ViT-FREE11_{FT}Exit Depth=11, Fine-tuning Protocol=Fine-tuned, GFLOPs=10.552026.06 | 92.83 | — | |
| ArcFaceVenue=CVPR '19, PPFR=No2024.03 | 92.77 | — | |
| ArcFaceBackbone=ResNet-100, Resolution=112x1122021.03 | 92.72 | — | |
| ViT-FREE12_{FT}Exit Depth=12, Fine-tuning Protocol=Fine-tuned, GFLOPs=11.502026.06 | 92.7 | — | |
| ViT-FREE10_{FT}Exit Depth=10, Fine-tuning Protocol=Fine-tuned, GFLOPs=9.602026.06 | 92.67 | — | |
| MagFaceTraining Data=MS1MV2, Backbone=ResNet-50, Venue=CVPR212024.10 | 92.49 | — | |
| ResNet50Model=ResNet502020.06 | 92.433 | — | |
| ResNet50Bits=FP322022.06 | 92.43 | — |