List your company
ResearchFormerly WizWand

5-way 1-shot Image Classification on miniImageNet (meta-test)

62.64AccuracyMetaOptNet
28.590437.430246.2755.1098May 28, 2021Jun 22, 2021Jul 17, 2021Aug 11, 2021Sep 5, 2021Sep 30, 2021Oct 25, 2021
Updated 8mo ago

Evaluation results

MethodLinks
MetaOptNetBackbone=ResNet-122021.1062.64—
DSNBackbone=ResNet-122021.1062.64—
ProtoNet MTM SPSA-TrackBackbone=ResNet-12, M=22021.1061.94—
ProtoNet MTM SPSA-TrackBackbone=ResNet-122021.1061.33—
vFSLBackbone=ResNet-122021.1061.23—
MTLBackbone=ResNet-122021.1061.2—
ProtoNet MTM SPSABackbone=ResNet-122021.1060.24—
Wang et al.Backbone=ResNet-122021.1059.84—
CAMLBackbone=ResNet-122021.1059.23—
Shot-FreeBackbone=ResNet-122021.1059.04—
TADAMBackbone=ResNet-122021.1058.5—
AdaResNetBackbone=ResNet-122021.1056.88—
ProtoNet (reproduced)Backbone=ResNet-122021.1056.52—
Bauer et al.Backbone=ResNet-342021.1056.3—
SNAILBackbone=ResNet-152021.1055.71—
Gidaris et al.Backbone=ResNet-152021.1055.45—
RAP-ProtoNetBackbone=ResNet-102021.1053.64—
Matching NetworksBackbone=ResNet-182021.1052.91—
Relation NetworksBackbone=ResNet-182021.1052.48—
Chen et al.Backbone=ResNet-182021.1051.87—
PACMAMLmeta-training data size (mi)=80, setting=ANIL, backbone=4 convolutional layers2021.0550.10.9
MAMLBackbone=ResNet-182021.1049.61—
PACMAMLmeta-training data size (mi)=20, setting=ANIL, backbone=4 convolutional layers2021.0549.10.9
PACMAMLmeta-training data size (mi)=40, setting=ANIL, backbone=4 convolutional layers2021.0548.90.9
MAMLmeta-training data size (mi)=80, setting=ANIL, backbone=4 convolutional layers2021.0548.10.9
MAMLmeta-training data size (mi)=20, setting=ANIL, backbone=4 convolutional layers2021.05480.9
PACMAMLmeta-training data size (mi)=10, setting=ANIL, backbone=4 convolutional layers2021.0547.80.9
MAMLmeta-training data size (mi)=40, setting=ANIL, backbone=4 convolutional layers2021.0547.80.9
MAMLmeta-training data size (mi)=10, setting=ANIL, backbone=4 convolutional layers2021.0547.30.9
FOMAMLmeta-training data size (mi)=40, setting=ANIL, backbone=4 convolutional layers2021.0546.21
FOMAMLmeta-training data size (mi)=80, setting=ANIL, backbone=4 convolutional layers2021.0545.70.9
PACOHmeta-training data size (mi)=80, setting=ANIL, backbone=4 convolutional layers2021.0544.60.9
FOMAMLmeta-training data size (mi)=20, setting=ANIL, backbone=4 convolutional layers2021.0544.30.9
BMAMLmeta-training data size (mi)=80, setting=ANIL, backbone=4 convolutional layers2021.0544.20.9
FOMAMLmeta-training data size (mi)=10, setting=ANIL, backbone=4 convolutional layers2021.0541.80.9
PACOHmeta-training data size (mi)=40, setting=ANIL, backbone=4 convolutional layers2021.0541.60.9
BMAMLmeta-training data size (mi)=40, setting=ANIL, backbone=4 convolutional layers2021.0541.50.9
PACOHmeta-training data size (mi)=20, setting=ANIL, backbone=4 convolutional layers2021.05370.9
BMAMLmeta-training data size (mi)=20, setting=ANIL, backbone=4 convolutional layers2021.0534.30.9
PACOHmeta-training data size (mi)=10, setting=ANIL, backbone=4 convolutional layers2021.0531.20.8
BMAMLmeta-training data size (mi)=10, setting=ANIL, backbone=4 convolutional layers2021.0529.90.9