5 datasets found

CIFAR-10-C and CIFAR-100-C are robustness benchmarks consisting of 19 corruptions types with five levels of severities.

CIFAR10-C and ImageNet-C datasets are used to evaluate the proposed framework for improving the natural robustness of trained DNNs against corrupted inputs.

CIFAR-10-C, CIFAR-100-C, and ImageNet-C are used as target datasets for corruption robustness evaluation.

Common corruptions and perturbations are used to evaluate the robustness of neural networks.

Benchmarking neural network robustness to common corruptions and perturbations.