Image Classification on SVHN 1,000 labels (train)
3.96Error Rate (%)Triple-GAN-V2
Evaluation results
| Method | Links | |
|---|---|---|
| Triple-GAN-V2Number of labels=1000, Data Augmentation=false, Classifier=13-layer CNN (classifier in [32])2019.12 | 3.96 | |
| SNTGNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 4.02 | |
| BadGANNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 4.25 | |
| VAT+EntNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 4.28 | |
| MMCVANumber of labels=1000, Data Augmentation=false, Extra Unlabeled Data=true2019.12 | 4.95 | |
| MTNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 5.21 | |
| MT (our code base)Number of labels=1000, Data Augmentation=false, Classifier=13-layer CNN (classifier in [32])2019.12 | 5.39 | |
| Π modelNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 5.73 | |
| Triple-GAN-V1Number of labels=1000, Data Augmentation=false, Classifier=13-layer CNN (classifier in [26])2019.12 | 5.77 | |
| VATNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 5.77 | |
| Pseudo LabelNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN2019.12 | 7.62 | |
| Improved-GANNumber of labels=1000, Data Augmentation=false, Classifier=13-layer CNN (classifier in [26])2019.12 | 8.11 | |
| SDGMNumber of labels=1000, Data Augmentation=false, Extra Unlabeled Data=true2019.12 | 16.61 | |
| ADGMNumber of labels=1000, Data Augmentation=false, Extra Unlabeled Data=true2019.12 | 22.86 | |
| M1+M2Number of labels=1000, Data Augmentation=false2019.12 | 36.02 |