Stacked What-Where Auto-encoders

Benchmark Model Rank Results
image-classification-on-cifar-10SWWAE#174Percentage correct: 92.2
image-classification-on-cifar-100SWWAE#165Percentage correct: 69.1
image-classification-on-mnistZhao et al. (2015) (auto-encoder)#47Percentage error: 4.76
image-classification-on-stl-10SWWAE#53Percentage correct: 74.3
semi-supervised-image-classification-on-stl-1SWWAE#10Accuracy: 74.30