Few-shot Image Classification on CelebA
87.1AccuracyOracle-MAML
Evaluation results
| Method | Links | |
|---|---|---|
| Oracle-MAMLway=2, shot=52018.10 | 87.1 | |
| Oracle-ProtoNetsway=2, shot=52018.10 | 85.13 | |
| BiGAN CACTUS-MAMLway=2, shot=52018.10 | 74.98 | |
| DeepCluster CACTUS-ProtoNetsway=2, shot=52018.10 | 74.15 | |
| DeepCluster CACTUS-MAMLway=2, shot=52018.10 | 73.79 | |
| BiGAN CACTUS-ProtoNetsway=2, shot=52018.10 | 65.58 | |
| Training from scratchway=2, shot=52018.10 | 63.19 | |
| DeepCluster knn-nearest neighborsway=2, shot=52018.10 | 61.47 | |
| DeepCluster MLP with dropoutway=2, shot=52018.10 | 60.65 | |
| DeepCluster linear classifierway=2, shot=52018.10 | 59.57 | |
| BiGAN linear classifierway=2, shot=52018.10 | 58.44 | |
| BiGAN MLP with dropoutway=2, shot=52018.10 | 56.26 | |
| BiGAN cluster matchingway=2, shot=52018.10 | 56.2 | |
| BiGAN knn-nearest neighborsway=2, shot=52018.10 | 56.15 | |
| DeepCluster cluster matchingway=2, shot=52018.10 | 51.51 |