# Face Recognition on CP-LFW Source: [Face Recognition on CP-LFW]() Collection index: [llms.txt]() Data updated: 2026-04-13T09:05:34.404Z Evaluation of face recognition performance on the CP-LFW dataset, typically measured using Accuracy. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Face Recognition | | Dataset (including any stated split) | CP-LFW | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 26 of 26 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | FaceLiVTv2-L | 93.07 | Not recorded | 69dc78697d0172140ec78986 | | FaceLiVTv2-M | 92.92 | Not recorded | 69dc78697d0172140ec78986 | | KANFace-0.5 | 92.65 | Not recorded | 69dc78697d0172140ec78986 | | EdgeFace-S | 92.56 | Not recorded | 69dc78697d0172140ec78986 | | FaceLiVTv2-S | 92.45 | Not recorded | 69dc78697d0172140ec78986 | | GhostFaceNetV1-1 | 91.93 | Not recorded | 69dc78697d0172140ec78986 | | KANFace-0.5 | 91.9 | Not recorded | 69dc78697d0172140ec78986 | | EdgeFace-XS | 91.82 | Not recorded | 69dc78697d0172140ec78986 | | EdgeFace-S | 91.63 | Not recorded | 69dc78697d0172140ec78986 | | KANFace-0.6 | 91.47 | Not recorded | 69dc78697d0172140ec78986 | | FaceLiVTv1-M | 90.97 | Not recorded | 69dc78697d0172140ec78986 | | SwiftFaceFormer-L3 | 90.7 | Not recorded | 69dc78697d0172140ec78986 | | FaceLiVTv1-S | 90.7 | Not recorded | 69dc78697d0172140ec78986 | | FaceLiVTv2-XS | 90.38 | Not recorded | 69dc78697d0172140ec78986 | | GhostFaceNetV2-2 | 90.17 | Not recorded | 69dc78697d0172140ec78986 | | SwiftFaceFormer-L1 | 90.1 | Not recorded | 69dc78697d0172140ec78986 | | GhostFaceNetV1-2 | 90.07 | Not recorded | 69dc78697d0172140ec78986 | | PocketNetM256 | 90.03 | Not recorded | 69dc78697d0172140ec78986 | | PocketNetM128 | 90 | Not recorded | 69dc78697d0172140ec78986 | | SwiftFaceFormer-S | 90 | Not recorded | 69dc78697d0172140ec78986 | | MobileFaceNet | 89.22 | Not recorded | 69dc78697d0172140ec78986 | | PocketNetS256 | 88.93 | Not recorded | 69dc78697d0172140ec78986 | | SwiftFaceFormer-XS | 88.65 | Not recorded | 69dc78697d0172140ec78986 | | PocketNetS128 | 88.63 | Not recorded | 69dc78697d0172140ec78986 | | VarGFaceNet | 88.55 | Not recorded | 69dc78697d0172140ec78986 | | ShuffleFaceNet-1.5 | 88.5 | Not recorded | 69dc78697d0172140ec78986 | ### FaceLiVTv2-L Accuracy: 93.07. Result ID: 69dcb1c947f34ce924c3d3b3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Year":"-","Param (M)":"8.52","FLOPs":"309M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.71(14.0× ↓)"}. This record does not assert independent reproduction. ### FaceLiVTv2-M Accuracy: 92.92. Result ID: 69dcb1c947f34ce924c3d3b4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Year":"-","Param (M)":"7.02","FLOPs":"258M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.65(16.7%↓)"}. This record does not assert independent reproduction. ### KANFace-0.5 Accuracy: 92.65. Result ID: 69dcb1c947f34ce924c3d3a1. [Reporting paper]() Recorded evaluation configuration: {"Year":"'25","Param (M)":"6.80","FLOPs":"397M","Training Dataset":"WebFace12M","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"9.98"}. This record does not assert independent reproduction. ### EdgeFace-S Accuracy: 92.56. Result ID: 69dcb1c947f34ce924c3d3a4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"3.6","FLOPs":"306M","Training Dataset":"WebFace12M","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"9.89"}. This record does not assert independent reproduction. ### FaceLiVTv2-S Accuracy: 92.45. Result ID: 69dcb1c947f34ce924c3d3b5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Year":"-","Param (M)":"4.62","FLOPs":"179M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.54(30.8%↓)"}. This record does not assert independent reproduction. ### GhostFaceNetV1-1 Accuracy: 91.93. Result ID: 69dcb1c947f34ce924c3d3ac. [Reporting paper]() Recorded evaluation configuration: {"Year":"'23","Param (M)":"4.1","FLOPs":"216M","Training Dataset":"MS1MV3","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.78"}. This record does not assert independent reproduction. ### KANFace-0.5 Accuracy: 91.9. Result ID: 69dcb1c947f34ce924c3d3a2. [Reporting paper]() Recorded evaluation configuration: {"Year":"'25","Param (M)":"6.80","FLOPs":"397M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Retrained from scratch","Latency (ms) on iPhone 15 Pro":"9.98"}. This record does not assert independent reproduction. ### EdgeFace-XS Accuracy: 91.82. Result ID: 69dcb1c947f34ce924c3d3af. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"1.77","FLOPs":"154M","Training Dataset":"WebFace12M","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"5.82"}. This record does not assert independent reproduction. ### EdgeFace-S Accuracy: 91.63. Result ID: 69dcb1c947f34ce924c3d3a5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"3.6","FLOPs":"306M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Retrained from scratch","Latency (ms) on iPhone 15 Pro":"9.89"}. This record does not assert independent reproduction. ### KANFace-0.6 Accuracy: 91.47. Result ID: 69dcb1c947f34ce924c3d3ad. [Reporting paper]() Recorded evaluation configuration: {"Year":"'25","Param (M)":"4.74","FLOPs":"240M","Training Dataset":"WebFace12M","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"6.52"}. This record does not assert independent reproduction. ### FaceLiVTv1-M Accuracy: 90.97. Result ID: 69dcb1c947f34ce924c3d3a3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Year":"'25","Param (M)":"9.8","FLOPs":"386M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.67"}. This record does not assert independent reproduction. ### SwiftFaceFormer-L3 Accuracy: 90.7. Result ID: 69dcb1c947f34ce924c3d39d. [Reporting paper]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"28.0","FLOPs":"2.0G","Training Dataset":"MS1MV3","Model Scale":"Large-Scale","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"2.59"}. This record does not assert independent reproduction. ### FaceLiVTv1-S Accuracy: 90.7. Result ID: 69dcb1c947f34ce924c3d3ae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Year":"'25","Param (M)":"5.89","FLOPs":"237M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.47"}. This record does not assert independent reproduction. ### FaceLiVTv2-XS Accuracy: 90.38. Result ID: 69dcb1c947f34ce924c3d3b6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Year":"-","Param (M)":"2.9","FLOPs":"90M","Training Dataset":"Glint360K","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.43(35.8%↓)"}. This record does not assert independent reproduction. ### GhostFaceNetV2-2 Accuracy: 90.17. Result ID: 69dcb1c947f34ce924c3d3b1. [Reporting paper]() Recorded evaluation configuration: {"Year":"'23","Param (M)":"6.8","FLOPs":"77M","Training Dataset":"MS1MV3","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.67"}. This record does not assert independent reproduction. ### SwiftFaceFormer-L1 Accuracy: 90.1. Result ID: 69dcb1c947f34ce924c3d3a6. [Reporting paper]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"11.8","FLOPs":"805M","Training Dataset":"MS1MV3","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"1.20"}. This record does not assert independent reproduction. ### GhostFaceNetV1-2 Accuracy: 90.07. Result ID: 69dcb1c947f34ce924c3d3b2. [Reporting paper]() Recorded evaluation configuration: {"Year":"'23","Param (M)":"4.1","FLOPs":"60M","Training Dataset":"MS1MV3","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.60"}. This record does not assert independent reproduction. ### PocketNetM256 Accuracy: 90.03. Result ID: 69dcb1c947f34ce924c3d39f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'22","Param (M)":"1.75","FLOPs":"1.1G","Training Dataset":"CASIA-WF","Model Scale":"Large-Scale","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.98"}. This record does not assert independent reproduction. ### PocketNetM128 Accuracy: 90. Result ID: 69dcb1c947f34ce924c3d3a0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'22","Param (M)":"1.68","FLOPs":"1.1G","Training Dataset":"CASIA-WF","Model Scale":"Large-Scale","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.98"}. This record does not assert independent reproduction. ### SwiftFaceFormer-S Accuracy: 90. Result ID: 69dcb1c947f34ce924c3d3a7. [Reporting paper]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"6.0","FLOPs":"485M","Training Dataset":"MS1MV3","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.65"}. This record does not assert independent reproduction. ### MobileFaceNet Accuracy: 89.22. Result ID: 69dcb1c947f34ce924c3d3a8. [Reporting paper]() Recorded evaluation configuration: {"Year":"'21","Param (M)":"0.99","FLOPs":"440M","Training Dataset":"MS1MV2","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.77"}. This record does not assert independent reproduction. ### PocketNetS256 Accuracy: 88.93. Result ID: 69dcb1c947f34ce924c3d3ab. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'22","Param (M)":"0.99","FLOPs":"587M","Training Dataset":"CASIA-WF","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.88"}. This record does not assert independent reproduction. ### SwiftFaceFormer-XS Accuracy: 88.65. Result ID: 69dcb1c947f34ce924c3d3b0. [Reporting paper]() Recorded evaluation configuration: {"Year":"'24","Param (M)":"3.4","FLOPs":"294M","Training Dataset":"MS1MV3","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.59"}. This record does not assert independent reproduction. ### PocketNetS128 Accuracy: 88.63. Result ID: 69dcb1c947f34ce924c3d3aa. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'22","Param (M)":"0.92","FLOPs":"587M","Training Dataset":"CASIA-WF","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.88"}. This record does not assert independent reproduction. ### VarGFaceNet Accuracy: 88.55. Result ID: 69dcb1c947f34ce924c3d39e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Year":"'19","Param (M)":"5.0","FLOPs":"1.0G","Training Dataset":"MS1MV3","Model Scale":"Large-Scale","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.83"}. This record does not assert independent reproduction. ### ShuffleFaceNet-1.5 Accuracy: 88.5. Result ID: 69dcb1c947f34ce924c3d3a9. [Reporting paper]() Recorded evaluation configuration: {"Year":"'21","Param (M)":"2.6","FLOPs":"577M","Training Dataset":"MS1MV2","Model Scale":"Lightweight","Experimental Setup":"Original","Latency (ms) on iPhone 15 Pro":"0.69"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.