Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/PocketNetS

PocketNetS

Reported on 48 benchmarks across 6 tasks · 1 paper · 18 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision16 results

Face ReconstructiononLFW
MFLOPs· 2021-08-24
587.24
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononCALFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononCPLFW
Accuracy· 2021-08-24
0.8893
best: 93.37 (DiscFace)
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononLFW
MFLOPs· 2021-08-24
587.24
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononCALFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononCPLFW
Accuracy· 2021-08-24
0.8893
best: 93.37 (DiscFace)
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononAgeDB-30
Accuracy· 2021-08-24
0.9635
best: 71.62 (USynthFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononCFP-FP
Accuracy· 2021-08-24
0.9334
best: 0.9951 (PartialFC (R200))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononLFW
Accuracy· 2021-08-24
0.9966
best: 0.998667 (GhostFaceNetV2-1 (MS1MV3))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononLFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face ReconstructiononCALFW
Accuracy· 2021-08-24
0.955
best: 96.15 (DiscFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononAgeDB-30
Accuracy· 2021-08-24
0.9635
best: 71.62 (USynthFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononCFP-FP
Accuracy· 2021-08-24
0.9334
best: 0.9951 (PartialFC (R200))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononLFW
Accuracy· 2021-08-24
0.9966
best: 0.998667 (GhostFaceNetV2-1 (MS1MV3))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononLFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ReconstructiononCALFW
Accuracy· 2021-08-24
0.955
best: 96.15 (DiscFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710

Methodology16 results

Face RecognitiononLFW
MFLOPs· 2021-08-24
587.24
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononCALFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononCPLFW
Accuracy· 2021-08-24
0.8893
best: 0.9465 (GhostFaceNetV2-1)
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonLFW
MFLOPs· 2021-08-24
587.24
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonCALFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonCPLFW
Accuracy· 2021-08-24
0.8893
best: 93.37 (DiscFace)
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononAgeDB-30
Accuracy· 2021-08-24
0.9635
best: 71.62 (USynthFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononCFP-FP
Accuracy· 2021-08-24
0.9334
best: 0.9933 (GhostFaceNetV2-1)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononLFW
Accuracy· 2021-08-24
0.9966
best: 0.998667 (GhostFaceNetV2-1 (MS1MV3))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononLFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Face RecognitiononCALFW
Accuracy· 2021-08-24
0.955
best: 0.96233 (Prodpoly)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonAgeDB-30
Accuracy· 2021-08-24
0.9635
best: 71.62 (USynthFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonCFP-FP
Accuracy· 2021-08-24
0.9334
best: 0.9951 (PartialFC (R200))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonLFW
Accuracy· 2021-08-24
0.9966
best: 0.998667 (GhostFaceNetV2-1 (MS1MV3))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonLFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3DonCALFW
Accuracy· 2021-08-24
0.955
best: 96.15 (DiscFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710

Music8 results

Facial Recognition and ModellingonLFW
MFLOPs· 2021-08-24
587.24
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonCALFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonCPLFW
Accuracy· 2021-08-24
0.8893
best: 93.37 (DiscFace)
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonAgeDB-30
Accuracy· 2021-08-24
0.9635
best: 71.62 (USynthFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonCFP-FP
Accuracy· 2021-08-24
0.9334
best: 0.9951 (PartialFC (R200))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonLFW
Accuracy· 2021-08-24
0.9966
best: 0.998667 (GhostFaceNetV2-1 (MS1MV3))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonLFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
Facial Recognition and ModellingonCALFW
Accuracy· 2021-08-24
0.955
best: 96.15 (DiscFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710

Medical8 results

3D Face ModellingonLFW
MFLOPs· 2021-08-24
587.24
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonCALFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonCPLFW
Accuracy· 2021-08-24
0.8893
best: 93.37 (DiscFace)
SOTA
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonAgeDB-30
Accuracy· 2021-08-24
0.9635
best: 71.62 (USynthFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonCFP-FP
Accuracy· 2021-08-24
0.9334
best: 0.9951 (PartialFC (R200))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonLFW
Accuracy· 2021-08-24
0.9966
best: 0.998667 (GhostFaceNetV2-1 (MS1MV3))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonLFW
MParams· 2021-08-24
0.99
best: 3.65 (EdgeFace - S (g=0.5))
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710
3D Face ModellingonCALFW
Accuracy· 2021-08-24
0.955
best: 96.15 (DiscFace)
PocketNet: Extreme Lightweight Face Recognition Network using Neural Architecture Search and Multi-Step Knowledge Distillation arXiv:2108.10710