5-way Few-shot Classification on miniImageNet standard (test)
71.55AccuracyGAP
Evaluation results
| Method | Links | |
|---|---|---|
| GAPBackbone=Conv-4, Shots=5-shot2023.04 | 71.55 | |
| Approximate GAPBackbone=Conv-4, Shots=5-shot2023.04 | 70.75 | |
| METALBackbone=Conv-4, Shots=5-shot2023.04 | 70.52 | |
| PAMELA†Backbone=Conv-4, Shots=5-shot2023.04 | 70.51 | |
| CxGradBackbone=Conv-4, Shots=5-shot2023.04 | 69.82 | |
| L2FBackbone=Conv-4, Shots=5-shot2023.04 | 69.38 | |
| ModGrad†Backbone=Conv-4, Shots=5-shot2023.04 | 69.17 | |
| ALFABackbone=Conv-4, Shots=5-shot2023.04 | 69.12 | |
| MCshot=5, backbone=ConvNet, data augmentation=true2021.10 | 68.47 | |
| WarpGradBackbone=Conv-4, Shots=5-shot2023.04 | 68.4 | |
| EEMLBackbone=Conv-4, Shots=5-shot2023.04 | 68.4 | |
| MAML++Backbone=Conv-4, Shots=5-shot2023.04 | 68.32 | |
| sparse-MAML+shot=5, backbone=ConvNet2021.10 | 68.05 | |
| MC1†Backbone=Conv-4, Shots=5-shot2023.04 | 68.01 | |
| MC2†Backbone=Conv-4, Shots=5-shot2023.04 | 67.99 | |
| MH-O†Backbone=Conv-4, Shots=5-shot2023.04 | 67.76 | |
| MH+Backbone=Conv-4, Shots=5-shot2023.04 | 67.16 | |
| sparse-MAMLshot=5, backbone=ConvNet2021.10 | 67.03 | |
| Sparse-MAML†Backbone=Conv-4, Shots=5-shot2023.04 | 67.03 | |
| Sparse-MAML++Backbone=Conv-4, Shots=5-shot2023.04 | 67.03 | |
| BOILBackbone=Conv-4, Shots=5-shot2023.04 | 66.46 | |
| BOILshot=5, backbone=ConvNet2021.10 | 66.45 | |
| HyperMAMLBackbone=Conv-4, Shots=5-shot2023.04 | 66.29 | |
| ECMLBackbone=Conv-4, Shots=5-shot2023.04 | 65.26 | |
| exp-MAMLshot=5, backbone=ConvNet2021.10 | 65.21 | |
| Sharp-MAML_bothBackbone=Conv-4, Shots=5-shot2023.04 | 65.04 | |
| sparse-ReLU-MAMLshot=5, backbone=ConvNet2021.10 | 64.84 | |
| Sparse-ReLU-MAML†Backbone=Conv-4, Shots=5-shot2023.04 | 64.84 | |
| FBMBackbone=Conv-4, Shots=5-shot2023.04 | 64.78 | |
| MAMLBackbone=Conv-4, Shots=5-shot2023.04 | 64.59 | |
| MH-C†Backbone=Conv-4, Shots=5-shot2023.04 | 64.52 | |
| BMAMLBackbone=Conv-4, Shots=5-shot2023.04 | 64.23 | |
| CTMLBackbone=Conv-4, Shots=5-shot2023.04 | 64.15 | |
| ARMLBackbone=Conv-4, Shots=5-shot2023.04 | 64.12 | |
| Meta-SGDshot=5, backbone=ConvNet2021.10 | 64.03 | |
| Meta-SGD†Backbone=Conv-4, Shots=5-shot2023.04 | 64 | |
| Sharp-MAML_lowBackbone=Conv-4, Shots=5-shot2023.04 | 63.18 | |
| MAMLshot=5, backbone=ConvNet2021.10 | 63.15 | |
| Sharp-MAML_upBackbone=Conv-4, Shots=5-shot2023.04 | 63.06 | |
| ANILshot=5, backbone=ConvNet2021.10 | 61.5 | |
| ANILBackbone=Conv-4, Shots=5-shot2023.04 | 61.5 | |
| SignMAMLBackbone=Conv-4, Shots=5-shot2023.04 | 60.7 | |
| GAPBackbone=Conv-4, Shots=1-shot2023.04 | 54.86 | |
| MCshot=1, backbone=ConvNet, data augmentation=true2021.10 | 54.23 | |
| MC2†Backbone=Conv-4, Shots=1-shot2023.04 | 54.08 | |
| BMAMLBackbone=Conv-4, Shots=1-shot2023.04 | 53.8 | |
| MC1†Backbone=Conv-4, Shots=1-shot2023.04 | 53.74 | |
| Approximate GAPBackbone=Conv-4, Shots=1-shot2023.04 | 53.52 | |
| PAMELA†Backbone=Conv-4, Shots=1-shot2023.04 | 53.5 | |
| ModGrad†Backbone=Conv-4, Shots=1-shot2023.04 | 53.2 | |
| METALBackbone=Conv-4, Shots=1-shot2023.04 | 52.63 | |
| MH-O†Backbone=Conv-4, Shots=1-shot2023.04 | 52.5 | |
| EEMLBackbone=Conv-4, Shots=1-shot2023.04 | 52.42 | |
| WarpGradBackbone=Conv-4, Shots=1-shot2023.04 | 52.3 | |
| MAML++Backbone=Conv-4, Shots=1-shot2023.04 | 52.15 | |
| L2FBackbone=Conv-4, Shots=1-shot2023.04 | 52.1 | |
| HyperMAMLBackbone=Conv-4, Shots=1-shot2023.04 | 51.84 | |
| CxGradBackbone=Conv-4, Shots=1-shot2023.04 | 51.8 | |
| MT-netshot=1, backbone=ConvNet2021.10 | 51.7 | |
| MT-netBackbone=Conv-4, Shots=1-shot2023.04 | 51.7 | |
| sparse-MAML+shot=1, backbone=ConvNet2021.10 | 51.04 | |
| Sparse-MAML++Backbone=Conv-4, Shots=1-shot2023.04 | 51.04 | |
| T-netBackbone=Conv-4, Shots=1-shot2023.04 | 50.86 | |
| FBMBackbone=Conv-4, Shots=1-shot2023.04 | 50.62 | |
| ALFABackbone=Conv-4, Shots=1-shot2023.04 | 50.58 | |
| Meta-SGDshot=1, backbone=ConvNet2021.10 | 50.47 | |
| Meta-SGD†Backbone=Conv-4, Shots=1-shot2023.04 | 50.47 | |
| CTMLBackbone=Conv-4, Shots=1-shot2023.04 | 50.47 | |
| ARMLBackbone=Conv-4, Shots=1-shot2023.04 | 50.42 | |
| sparse-ReLU-MAMLshot=1, backbone=ConvNet2021.10 | 50.39 | |
| Sparse-ReLU-MAML†Backbone=Conv-4, Shots=1-shot2023.04 | 50.39 | |
| sparse-MAMLshot=1, backbone=ConvNet2021.10 | 50.35 | |
| Sparse-MAML†Backbone=Conv-4, Shots=1-shot2023.04 | 50.35 | |
| Sharp-MAML_bothBackbone=Conv-4, Shots=1-shot2023.04 | 50.28 | |
| PLATIPUSBackbone=Conv-4, Shots=1-shot2023.04 | 50.13 | |
| Sharp-MAML_lowBackbone=Conv-4, Shots=1-shot2023.04 | 49.72 | |
| BOILshot=1, backbone=ConvNet2021.10 | 49.61 | |
| BOILBackbone=Conv-4, Shots=1-shot2023.04 | 49.61 | |
| Sharp-MAML_upBackbone=Conv-4, Shots=1-shot2023.04 | 49.56 | |
| MH+Backbone=Conv-4, Shots=1-shot2023.04 | 49.41 | |
| LLAMABackbone=Conv-4, Shots=1-shot2023.04 | 49.4 | |
| IMAML-HFBackbone=Conv-4, Shots=1-shot2023.04 | 49.3 | |
| ECMLBackbone=Conv-4, Shots=1-shot2023.04 | 48.94 | |
| MH-C†Backbone=Conv-4, Shots=1-shot2023.04 | 48.64 | |
| exp-MAMLshot=1, backbone=ConvNet2021.10 | 48.38 | |
| MAMLshot=1, backbone=ConvNet2021.10 | 48.07 | |
| MAMLBackbone=Conv-4, Shots=1-shot2023.04 | 47.89 | |
| Shrinkageshot=1, backbone=ConvNet2021.10 | 47.7 | |
| ANILshot=1, backbone=ConvNet2021.10 | 46.7 | |
| ANILBackbone=Conv-4, Shots=1-shot2023.04 | 46.7 | |
| SignMAMLBackbone=Conv-4, Shots=1-shot2023.04 | 42.9 |