CNN Filter DB: An Empirical Investigation of Trained Convolutional Filters

Open paper
Benchmark Model Rank Results
image-classification-on-cifar-10ResNet-9#142Percentage correct: 94.79
image-classification-on-cifar-100ResNet-9#144Percentage correct: 75.59
image-classification-on-fashion-mnistInception v3#5Percentage error: 5.56Accuracy: 94.44
image-classification-on-kuzushiji-mnistResNet-14#6Accuracy: 98.75
image-classification-on-mnistResNet-9#50Accuracy: 99.68

Rank counts only results with a code link.