Image Classification on CIFAR-10 1% BER (test)
92.6Clean AccuracyResNet18
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ResNet18Generator=None2023.06 | 92.6 | 38.9 | — | |
| ResNet50Generator=None2023.06 | 92.6 | 26.1 | — | |
| VGG19Generator=None2023.06 | 90.5 | 36 | — | |
| VGG16Generator=None2023.06 | 90.3 | 35.7 | — | |
| NeuralFuseBase Model=VGG16, Generator=ConvL2023.06 | 90.1 | 86 | 50.3 | |
| NeuralFuseBase Model=ResNet18, Generator=ConvL2023.06 | 89.8 | 87.8 | 48.8 | |
| NeuralFuseBase Model=VGG19, Generator=ConvL2023.06 | 89.8 | 77.7 | 41.7 | |
| NeuralFuseBase Model=VGG16, Generator=DeConvL2023.06 | 89.7 | 85.5 | 49.8 | |
| NeuralFuseBase Model=ResNet18, Generator=DeConvL2023.06 | 89.6 | 88.5 | 49.6 | |
| NeuralFuseBase Model=VGG11, Generator=ConvL2023.06 | 89.6 | 87.2 | 45.1 | |
| NeuralFuseBase Model=VGG11, Generator=DeConvL2023.06 | 89.3 | 87.2 | 45 | |
| VGG11Generator=None2023.06 | 88.4 | 42.2 | — | |
| NeuralFuseBase Model=ResNet18, Generator=ConvS2023.06 | 88.2 | 59.5 | 20.6 | |
| NeuralFuseBase Model=ResNet50, Generator=DeConvL2023.06 | 87.4 | 63.3 | 37.2 | |
| NeuralFuseBase Model=VGG16, Generator=ConvS2023.06 | 87.4 | 59.6 | 23.9 | |
| NeuralFuseBase Model=VGG16, Generator=UNetL2023.06 | 87.4 | 83.4 | 47.7 | |
| NeuralFuseBase Model=VGG16, Generator=UNetS2023.06 | 87.4 | 71.2 | 35.5 | |
| NeuralFuseBase Model=VGG19, Generator=ConvS2023.06 | 87.3 | 52.7 | 16.7 | |
| NeuralFuseBase Model=VGG11, Generator=UNetL2023.06 | 87.1 | 83.6 | 41.4 | |
| NeuralFuseBase Model=VGG16, Generator=DeConvS2023.06 | 86.8 | 66.5 | 30.8 | |
| NeuralFuseBase Model=ResNet18, Generator=UNetL2023.06 | 86.6 | 84.6 | 45.6 | |
| NeuralFuseBase Model=VGG19, Generator=DeConvS2023.06 | 86.5 | 58.2 | 22.2 | |
| NeuralFuseBase Model=VGG19, Generator=DeConvL2023.06 | 86.3 | 78.4 | 42.4 | |
| NeuralFuseBase Model=VGG19, Generator=UNetL2023.06 | 86.3 | 82.1 | 46 | |
| NeuralFuseBase Model=VGG19, Generator=UNetS2023.06 | 86.3 | 66.4 | 30.4 | |
| NeuralFuseBase Model=ResNet50, Generator=UNetL2023.06 | 86.2 | 75.5 | 49.4 | |
| NeuralFuseBase Model=VGG11, Generator=DeConvS2023.06 | 85.6 | 68.2 | 26 | |
| NeuralFuseBase Model=ResNet50, Generator=ConvL2023.06 | 85.5 | 53.2 | 27.1 | |
| NeuralFuseBase Model=VGG11, Generator=UNetS2023.06 | 85.5 | 72.7 | 30.5 | |
| NeuralFuseBase Model=ResNet50, Generator=ConvS2023.06 | 85.2 | 34.6 | 8.5 | |
| NeuralFuseBase Model=VGG11, Generator=ConvS2023.06 | 84.9 | 66.3 | 24.1 | |
| NeuralFuseBase Model=ResNet18, Generator=UNetS2023.06 | 84.4 | 68.8 | 29.8 | |
| NeuralFuseBase Model=ResNet18, Generator=DeConvS2023.06 | 82.9 | 68.8 | 29.9 | |
| NeuralFuseBase Model=ResNet50, Generator=DeConvS2023.06 | 82.4 | 42.2 | 16.1 | |
| NeuralFuseBase Model=ResNet50, Generator=UNetS2023.06 | 77.3 | 56.2 | 30.1 |