CIFAR-10-C

CIFAR-10-C is a benchmark dataset that contains corrupted versions of the CIFAR-10 dataset, used to evaluate the robustness of models against various types of distribution shifts and adversarial examples during test-time adaptation.

Papers using CIFAR-10-C (1)

Medbn: Robust Test-time Adaptation Against Malicious Test Samples2024 · 2 cites