Image Classification on CIFAR-10 (6-way sharded)
41.3AccuracyEnsemble
Evaluation results
| Method | Links | |
|---|---|---|
| EnsembleBackbone=VGG11, Fusion data points=4002025.06 | 41.3 | |
| KF-Linear ConductanceBackbone=VGG11, Fusion data points=4002025.06 | 40.3 | |
| KF-Linear DeepLIFTBackbone=VGG11, Fusion data points=4002025.06 | 40.1 | |
| KF-Linear UniformBackbone=VGG11, Fusion data points=4002025.06 | 35.8 | |
| KF-Gradient ConductanceBackbone=VGG11, Fusion data points=4002025.06 | 34.8 | |
| KF-Gradient DeepLIFTBackbone=VGG11, Fusion data points=4002025.06 | 34.7 | |
| KF-Gradient UniformBackbone=VGG11, Fusion data points=4002025.06 | 33.8 | |
| Individual ModelsBackbone=VGG11, Fusion data points=4002025.06 | 19.9 | |
| KDBackbone=VGG11, Fusion data points=400, Initialization strategy=randomly initialized2025.06 | 17.4 | |
| LPBackbone=VGG11, Fusion data points=400, Initialization strategy=weights of the first base model2025.06 | 13.3 | |
| Vanilla AveragingBackbone=VGG11, Fusion data points=4002025.06 | 10 | |
| OTF UniformBackbone=VGG11, Fusion data points=4002025.06 | 10 | |
| OTF ConductanceBackbone=VGG11, Fusion data points=4002025.06 | 10 | |
| OTF DeepLIFTBackbone=VGG11, Fusion data points=4002025.06 | 10 |