List your company
ResearchFormerly WizWand

Few-shot classification on Places

84.41AccuracySupervised
19.46236.323553.18570.0465Jan 23, 2020Feb 11, 2021Mar 3, 2022Mar 23, 2023Apr 11, 2024May 1, 2025May 21, 2026
Updated 5mo ago

Evaluation results

MethodLinks
SupervisedBackbone=ResNet-50, Evaluation protocol=5-way 5-shot2023.0384.41——
PsCoBackbone=ResNet-50, Base Method=MoCo v2, Evaluation protocol=5-way 5-shot2023.0383.87——
PsCoBackbone=ResNet-50, Base Method=BYOL, Evaluation protocol=5-way 5-shot2023.0383.8——
MoCo v2Backbone=ResNet-50, Evaluation protocol=5-way 5-shot2023.0381.39——
PsCoBackbone=ResNet-18, Base Method=MoCo v2, Evaluation protocol=5-way 5-shot2023.0379.74——
SRasPBackbone=ResNet-10, FT=True, Shot=5-shot2026.0379.4——
StyleAdvBackbone=ResNet-10, FT=True, Shot=5-shot2026.0379.35——
MoCo v2Backbone=ResNet-18, Evaluation protocol=5-way 5-shot2023.0379.11——
SVasPBackbone=ResNet-10, FT=True, Shot=5-shot2026.0378.91——
SRasPBackbone=ResNet-10, FT=False, Shot=5-shot2026.0378.57——
HAPBackbone=ResNet-10, FT=True, Shot=5-shot2026.0378.46——
HAPBackbone=ResNet-10, FT=True, Shot=5-shot2026.0377.84——
SVasPBackbone=ResNet-10, FT=False, Shot=5-shot2026.0377.78——
StyleAdvBackbone=ResNet-10, FT=False, Shot=5-shot2026.0377.73——
ATABackbone=ResNet-10, FT=True, Shot=5-shot2026.0376.64——
GNN w/ AFA (Ours)Shot=5-shot2022.0876.21——
ATABackbone=ResNet-10, FT=False, Shot=5-shot2026.0375.48——
BYOLBackbone=ResNet-50, Evaluation protocol=5-way 5-shot2023.0375.43——
RDCBackbone=ResNet-10, FT=True, Shot=5-shot2026.0374.65——
PsCoway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0373.95——
GNN (FT)Backbone=ResNet-10, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=true2020.0173.94——
FWTBackbone=ResNet-10, FT=False, Shot=5-shot2026.0373.94——
TPN w/ AFA (Ours)Shot=5-shot2022.0872.81——
FLoRBackbone=ResNet-10, FT=True, Shot=5-shot2026.0372.37——
FLoRBackbone=ResNet-10, FT=False, Shot=5-shot2026.0372.31——
NSAEBackbone=ResNet-10, FT=True, Shot=5-shot2026.0371.02——
GNNBackbone=ResNet-10, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=false2020.0170.84——
GNNBackbone=ResNet-10, FT=False, Shot=5-shot2026.0370.84——
Fine-tuneBackbone=ResNet-10, FT=True, Shot=5-shot2026.0370.06——
SimCLRway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0369.93——
SWAVway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0369.7——
MoCo v2way=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0369.08——
MN w/ AFA (Ours)Shot=5-shot2022.0868.87——
ProtoNetsway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0367.77——
MAMLway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0367.54——
RelationNet (FT)Backbone=ResNet-10, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=true2020.0166.28——
MetaOptNet-SVM-trainvalBackbone=ResNet-12, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=false2020.0165.83——
Meta-GMVAEway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0365.08——
MatchingNet (FT)Backbone=ResNet-10, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=true2020.0164.55——
PsCoway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0363.6——
RelationNetBackbone=ResNet-10, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=false2020.0163.32——
MatchingNetBackbone=ResNet-10, Training Dataset=mini-ImageNet, FT (Feature-wise Transformation)=false2020.0163.16——
Meta-SVEBMway=5, shot=20, Backbone=Conv5, Pre-trained=miniImageNet2023.0361.09——
TPN w/ AFA (Ours)Shot=1-shot2022.0860.29——
SRasPBackbone=ResNet-10, FT=False, Shot=1-shot2026.0360.11——
SRasPBackbone=ResNet-10, FT=True, Shot=1-shot2026.0360.11——
SimCLRway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0360.1——
MAMLway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0360.05——
SWAVway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0359.57——
ProtoNetsway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0359.39——
MoCo v2way=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0359.09——
SVasPBackbone=ResNet-10, FT=False, Shot=1-shot2026.0359.07——
SVasPBackbone=ResNet-10, FT=True, Shot=1-shot2026.0359.07——
HAPBackbone=ResNet-10, FT=True, Shot=1-shot2026.0359.02——
StyleAdvBackbone=ResNet-10, FT=False, Shot=1-shot2026.0358.58——
StyleAdvBackbone=ResNet-10, FT=True, Shot=1-shot2026.0358.58——
HAPBackbone=ResNet-10, FT=False, Shot=1-shot2026.0357.71——
Meta-GMVAEway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0357.7——
FWTBackbone=ResNet-10, FT=False, Shot=1-shot2026.0355.77——
MN w/ AFA (Ours)Shot=1-shot2022.0854.66——
ATABackbone=ResNet-10, FT=True, Shot=1-shot2026.0354.18——
GNN w/ AFA (Ours)Shot=1-shot2022.0854.04——
FLoRBackbone=ResNet-10, FT=True, Shot=1-shot2026.0353.61——
ATABackbone=ResNet-10, FT=False, Shot=1-shot2026.0353.57——
FLoRBackbone=ResNet-10, FT=False, Shot=1-shot2026.0353.18——
GNNBackbone=ResNet-10, FT=False, Shot=1-shot2026.0353.1——
Meta-SVEBMway=5, shot=5, Backbone=Conv5, Pre-trained=miniImageNet2023.0351.27——
Weight RecyclingK=42026.0541.98——
NCCK=42026.0538.71——
Weight RecyclingK=22026.0535.94——
Weight RecyclingK=12026.0535.51——
NCCK=22026.0530.22——
KNNK=42026.0528.77——
KNNK=22026.0522.49——
NCCK=12026.0521.96——
KNNK=12026.0521.96——
GNNTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—52.4270.91
GNN + ATATraining dataset=mini-ImageNet, Number of ways=5-way2021.04—53.5775.48
GNN + FTTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—53.4470.7
GNN + LRPTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—52.2871.38
RelationNetTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—48.1664.25
RelationNet + ATATraining dataset=mini-ImageNet, Number of ways=5-way2021.04—51.1666.9
RelationNet + FTTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—49.9265.55
RelationNet + LRPTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—48.4765.35
TPNTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—56.1771.39
TPN + ATATraining dataset=mini-ImageNet, Number of ways=5-way2021.04—57.0372.12
TPN + FTTraining dataset=mini-ImageNet, Number of ways=5-way2021.04—52.4566.75