EnAET: A Self-Trained framework for Semi-Supervised and Supervised Learning with Ensemble Transformations

Benchmark Model Rank Results
semi-supervised-image-classification-on-svhn-1000-labelsEnAET#4Accuracy: 97.58
semi-supervised-image-classification-on-svhn-250-labelsEnAET#5Accuracy: 96.79
semi-supervised-image-classification-on-cifar-10-4000-labelsEnAET#6Percentage error: 4.18
semi-supervised-image-classification-on-stl-10-1000-labelsEnAET#6Accuracy: 91.96
image-classification-on-stl-10EnAET#9Percentage correct: 95.48
semi-supervised-image-classification-on-cifar-100-10000-labelsEnAET (WRN-28-2-Large)#10Percentage error: 22.92
semi-supervised-image-classification-on-cifar-100-10000-labelsEnAET (WRN-28-2)#15Percentage error: 26.93±0.21
image-classification-on-svhnEnAET#26Percentage error: 2.22
image-classification-on-cifar-10EnAET#77Percentage correct: 98.01
image-classification-on-cifar-100EnAET#118Percentage correct: 83.13