Classification on Omniglot to EMNIST (test)
90.81AccuracyHyperShot
Evaluation Results
| Method | Links | |
|---|---|---|
| HyperShotn-shot=5-shot, finetuning=true2022.03 | 90.81 | |
| DKT + BNCosSimshots=5-shot2019.10 | 90.3 | |
| DKTn-shot=5-shot2022.03 | 90.3 | |
| DKT + Linearshots=5-shot2019.10 | 89.51 | |
| HyperShotn-shot=5-shot, finetuning=false2022.03 | 89.04 | |
| DKT + CosSimshots=5-shot2019.10 | 88.1 | |
| RelationNetshots=5-shot2019.10 | 87.84 | |
| RelationNetn-shot=5-shot2022.03 | 87.84 | |
| OVE PG GP + Cosine (PL)n-shot=5-shot2022.03 | 87.52 | |
| MatchingNetshots=5-shot2019.10 | 87.41 | |
| MatchingNetn-shot=5-shot2022.03 | 87.41 | |
| ProtoNetshots=5-shot2019.10 | 87.22 | |
| ProtoNetn-shot=5-shot2022.03 | 87.22 | |
| OVE PG GP + Cosine (ML)n-shot=5-shot2022.03 | 86.22 | |
| Feature Transfershots=5-shot2019.10 | 86.1 | |
| Feature Transfern-shot=5-shot2022.03 | 86.1 | |
| MAMLshots=5-shot2019.10 | 83.54 | |
| MAMLn-shot=5-shot2022.03 | 83.54 | |
| HyperShotn-shot=1-shot, finetuning=true2022.03 | 80.65 | |
| Baseline++shots=5-shot2019.10 | 80.01 | |
| Baseline++n-shot=5-shot2022.03 | 80.01 | |
| HyperShotn-shot=1-shot, finetuning=false2022.03 | 78.06 | |
| OVE PG GP + Cosine (PL)n-shot=1-shot2022.03 | 77 | |
| DKTKernel=Linear, Shots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 75.97 | |
| DKT + Linearshots=1-shot2019.10 | 75.97 | |
| RelationNetShots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 75.62 | |
| RelationNetshots=1-shot2019.10 | 75.62 | |
| RelationNetn-shot=1-shot2022.03 | 75.62 | |
| DKTKernel=BNCosSim, Shots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 75.4 | |
| DKT + BNCosSimshots=1-shot2019.10 | 75.4 | |
| DKTn-shot=1-shot2022.03 | 75.4 | |
| MatchingNetShots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 75.01 | |
| MatchingNetshots=1-shot2019.10 | 75.01 | |
| MatchingNetn-shot=1-shot2022.03 | 75.01 | |
| DKTKernel=CosSim, Shots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 73.06 | |
| DKT + CosSimshots=1-shot2019.10 | 73.06 | |
| MAMLShots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 72.68 | |
| MAMLshots=1-shot2019.10 | 72.68 | |
| MAMLn-shot=1-shot2022.03 | 72.68 | |
| ProtoNetShots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 72.04 | |
| ProtoNetshots=1-shot2019.10 | 72.04 | |
| ProtoNetn-shot=1-shot2022.03 | 72.04 | |
| OVE PG GP + Cosine (ML)n-shot=1-shot2022.03 | 68.43 | |
| Bayesian MAMLn-shot=5-shot2022.03 | 65.26 | |
| Feature TransferShots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 64.22 | |
| Feature Transfershots=1-shot2019.10 | 64.22 | |
| Feature Transfern-shot=1-shot2022.03 | 64.22 | |
| Bayesian MAMLn-shot=1-shot2022.03 | 63.94 | |
| Baseline++Shots=1-shot, Ways=5-ways, Backbone=Conv-4, Number of runs=32019.10 | 56.84 | |
| Baseline++shots=1-shot2019.10 | 56.84 | |
| Baseline++n-shot=1-shot2022.03 | 56.84 |