Few-shot classification on MiniImageNet 5-Shot 5-Way (test)
0.8205AccuracyFEAT
Evaluation results
| Method | Links | |
|---|---|---|
| FEATSetup=Embedding adaptation, Backbone=ResNet2018.12 | 0.8205 | |
| GCNSetup=Embedding adaptation, Backbone=ResNet2018.12 | 0.8165 | |
| DEEPSETSSetup=Embedding adaptation, Backbone=ResNet2018.12 | 0.8093 | |
| BILSTMSetup=Embedding adaptation, Backbone=ResNet2018.12 | 0.8062 | |
| ProtoNetSetup=Instance embedding, Backbone=ResNet2018.12 | 0.8053 | |
| CTMBackbone=ResNet2018.12 | 0.8051 | |
| SimpleShotBackbone=ResNet2018.12 | 0.8002 | |
| MetaOptNetBackbone=ResNet2018.12 | 0.7863 | |
| TADAMBackbone=ResNet2018.12 | 0.767 | |
| PFABackbone=ResNet2018.12 | 0.7374 | |
| sparse-ReLU-MAMLBackbone=ResNet-12, Number of seeds=32021.10 | 0.7301 | |
| FEATSetup=Embedding adaptation, Backbone=ConvNet2018.12 | 0.7161 | |
| ProtoNetSetup=Instance embedding, Backbone=ConvNet2018.12 | 0.7133 | |
| DEEPSETSSetup=Embedding adaptation, Backbone=ConvNet2018.12 | 0.7096 | |
| MetaOptNetBackbone=ResNet-12, Number of seeds=3, Regularization=None2021.10 | 0.7088 | |
| GCNSetup=Embedding adaptation, Backbone=ConvNet2018.12 | 0.7059 | |
| BOILBackbone=ResNet-12, Number of seeds=32021.10 | 0.705 | |
| ANILBackbone=ResNet-12, Number of seeds=32021.10 | 0.7003 | |
| sparse-MAMLBackbone=ResNet-12, Number of seeds=32021.10 | 0.7002 | |
| MAMLBackbone=ResNet-12, Number of seeds=32021.10 | 0.6936 | |
| BILSTMSetup=Embedding adaptation, Backbone=ConvNet2018.12 | 0.6915 | |
| ProtoNetBackbone=ConvNet2018.12 | 0.682 | |
| PFABackbone=ConvNet2018.12 | 0.6787 | |
| RelationNetBackbone=ConvNet2018.12 | 0.6707 | |
| SimpleShotBackbone=ConvNet2018.12 | 0.6692 | |
| MAMLBackbone=ConvNet2018.12 | 0.6311 | |
| MatchNetBackbone=ConvNet2018.12 | 0.5109 |