List your company
ResearchFormerly WizWand

5-way Image Classification on CIFAR FS (test)

86.1AccuracyGidaris et al.
53.75662.15370.5578.947Apr 7, 2019Jul 15, 2019Oct 23, 2019Jan 31, 2020May 9, 2020Aug 17, 2020Nov 25, 2020
Updated 8mo ago

Evaluation results

MethodLinks
Gidaris et al.Backbone=WRN-28-10, Shot=5-shot2020.0486.1——
SIBBackbone=WRN-28-10, Shot=5-shot, K=3, learning_rate=1e-32020.0485.3——
SIBBackbone=WRN-28-10, Shot=5-shot, K=5, learning_rate=1e-32020.0485.3——
MetaOptNet-SVM-trainvalbackbone=ResNet-12, shot=5-shot, training_set=meta-train + meta-validation2019.0485——
SIBBackbone=WRN-28-10, Shot=5-shot, K=1, learning_rate=1e-32020.0484.9——
MetaOptNet-RRbackbone=ResNet-12, shot=5-shot2019.0484.3——
MetaOptNet-RRBackbone=ResNet-12, Shot=5-shot2020.0484.3——
MetaOptNet-SVMbackbone=ResNet-12, shot=5-shot2019.0484.2——
MetaOptNet-SVMBackbone=ResNet-12, Shot=5-shot2020.0484.2——
ProtoNets (our backbone)backbone=ResNet-12, shot=5-shot2019.0483.5——
SIBBackbone=WRN-28-10, Shot=5-shot, K=0, Feature=Pre-trained2020.0483.5——
MTUNetNumber of shots=5-shot, Backbone=WRN-28-102020.1182.93——
MTUNetNumber of shots=5-shot, Backbone=ResNet-182020.1180.16——
SIBBackbone=WRN-28-10, Shot=1-shot, K=5, learning_rate=1e-32020.0480——
Gidaris et al.Backbone=Conv-4-64, Shot=5-shot2020.0479.8——
R2D2backbone=96-192-384-512, shot=5-shot2019.0479.4——
SIBBackbone=WRN-28-10, Shot=1-shot, K=3, learning_rate=1e-32020.0478.4——
R2-D2Number of shots=5-shot2020.1178.3——
R2-D2Backbone=Conv-4-64, Shot=5-shot2020.0477.4——
SIBBackbone=Conv-4-64, Shot=5-shot, K=3, learning_rate=1e-32020.0477.1——
SIBBackbone=WRN-28-10, Shot=1-shot, K=1, learning_rate=1e-32020.0476.8——
SIBBackbone=Conv-4-64, Shot=5-shot, K=0, Feature=Pre-trained2020.0475.4——
GNNBackbone=Conv-4-64, Shot=5-shot2020.0475.3——
GNNNumber of shots=5-shot2020.1175.3——
Gidaris et al.Backbone=WRN-28-10, Shot=1-shot2020.0473.6——
MetaOptNet-SVM-trainvalbackbone=ResNet-12, shot=1-shot, training_set=meta-train + meta-validation2019.0472.8——
MetaOptNet-RRbackbone=ResNet-12, shot=1-shot2019.0472.6——
MetaOptNet-RRBackbone=ResNet-12, Shot=1-shot2020.0472.6——
ProtoNets (our backbone)backbone=ResNet-12, shot=1-shot2019.0472.2——
Prototypical Networksbackbone=64-64-64-64, shot=5-shot2019.0472——
MetaOptNet-SVMbackbone=ResNet-12, shot=1-shot2019.0472——
Prototypical NetBackbone=Conv-4-64, Shot=5-shot2020.0472——
MetaOptNet-SVMBackbone=ResNet-12, Shot=1-shot2020.0472——
ProtoNetNumber of shots=5-shot2020.1172——
MAMLbackbone=32-32-32-32, shot=5-shot2019.0471.5——
MAMLBackbone=Conv-4-64, Shot=5-shot2020.0471.5——
MAMLNumber of shots=5-shot2020.1171.5——
SIBBackbone=WRN-28-10, Shot=1-shot, K=0, Feature=Pre-trained2020.0470——
Relation Networksbackbone=64-96-128-256, shot=5-shot2019.0469.3——
Relation NetBackbone=Conv-4-64, Shot=5-shot2020.0469.3——
RelationNetNumber of shots=5-shot2020.1169.3——
SIBBackbone=Conv-4-64, Shot=1-shot, K=3, learning_rate=1e-32020.0468.7——
MTUNetNumber of shots=1-shot, Backbone=WRN-28-102020.1168.34——
MTUNetNumber of shots=1-shot, Backbone=ResNet-182020.1166.31——
R2D2backbone=96-192-384-512, shot=1-shot2019.0465.3——
R2-D2Number of shots=1-shot2020.1165.3——
Gidaris et al.Backbone=Conv-4-64, Shot=1-shot2020.0463.5——
R2-D2Backbone=Conv-4-64, Shot=1-shot2020.0462.3——
GNNBackbone=Conv-4-64, Shot=1-shot2020.0461.9——
GNNNumber of shots=1-shot2020.1161.9——
SIBBackbone=Conv-4-64, Shot=1-shot, K=0, Feature=Pre-trained2020.0459.2——
MAMLbackbone=32-32-32-32, shot=1-shot2019.0458.9——
MAMLBackbone=Conv-4-64, Shot=1-shot2020.0458.9——
MAMLNumber of shots=1-shot2020.1158.9——
Prototypical Networksbackbone=64-64-64-64, shot=1-shot2019.0455.5——
Prototypical NetBackbone=Conv-4-64, Shot=1-shot2020.0455.5——
ProtoNetNumber of shots=1-shot2020.1155.5——
Relation Networksbackbone=64-96-128-256, shot=1-shot2019.0455——
Relation NetBackbone=Conv-4-64, Shot=1-shot2020.0455——
RelationNetNumber of shots=1-shot2020.1155——
Baseline++Backbone=WRN-28-10, N-way=5-way2019.07—67.580.08
DCOBackbone=WRN-28-10, N-way=5-way2019.07—7284.2
DPGNbackbone=ConvNet, transduction=true2020.03—76.488.4
DPGNbackbone=ResNet12, transduction=true2020.03—77.990.2
Fine-tuningArchitecture=WRN-28-10, Pre-training Split=train2019.09—68.7286.11
Fine-tuningArchitecture=WRN-28-10, Pre-training Split=train + val2019.09—70.0787.26
GNNBackbone=Conv4-64, Type=Metric2020.09—61.975.3
MAMLBackbone=WRN-28-10, N-way=5-way2019.07—58.971.5
MAMLbackbone=ConvNet2020.03—58.971.5
MAMLBackbone=Conv4-32, Type=Meta2020.09—58.971.5
Manifold MixupBackbone=WRN-28-10, N-way=5-way2019.07—69.283.42
MetaOpt SVMArchitecture=ResNet-12, Pre-training Split=train2019.09—7284.2
MetaOpt SVMArchitecture=ResNet-12*, Pre-training Split=train + val2019.09—72.885
MetaOptNetbackbone=ResNet122020.03—7284.2
ProtoNetBackbone=WRN-28-10, N-way=5-way2019.07—55.572
ProtoNetbackbone=ConvNet2020.03—55.572
PROTOTYPICAL NETSBackbone=Conv4-64, Type=Metric2020.09—55.572
R2-D2Backbone=Conv4-512, Type=Metric2020.09—65.379.4
R2D2backbone=ConvNet2020.03—65.379.4
R2D2Architecture=conv (96)x42019.09—65.479.4
RCNBackbone=Conv4-64, Type=Metric2020.09—61.6177.63
RCNBackbone=Res12, Type=Metric2020.09—69.0282.96
RELATION NETSBackbone=Conv4-64, Type=Metric2020.09—5569.3
RelationNetBackbone=WRN-28-10, N-way=5-way2019.07—5569.3
RelationNetbackbone=ConvNet2020.03—5569.3
RotationBackbone=WRN-28-10, N-way=5-way2019.07—70.6684.15
S2M2_RBackbone=WRN-28-10, N-way=5-way2019.07—74.8187.47
S2M2_RBackbone=ResNet-18, N-way=5-way2019.07—63.6676.07
S2M2_RBackbone=ResNet-34, N-way=5-way2019.07—62.7775.75
Shot-Freebackbone=ResNet122020.03—69.284.7
Support-based initializationArchitecture=WRN-28-10, Pre-training Split=train2019.09—70.2683.82
Support-based initializationArchitecture=WRN-28-10, Pre-training Split=train + val2019.09—72.1485.21
Transductive fine-tuningArchitecture=WRN-28-10, Pre-training Split=train2019.09—76.5885.79
Transductive fine-tuningArchitecture=WRN-28-10, Pre-training Split=train + val2019.09—78.3687.54