CIFAR-10-C

'CIFAR-10-C' is a benchmark dataset that contains corrupted versions of the CIFAR-10 images and is used to evaluate the robustness of machine learning models under various types of noise and distortions.

Papers using CIFAR-10-C (1)

PerAda: Parameter-Efficient Federated Learning Personalization with Generalization Guarantees2023 · 3 cites