Semantic-Discrepant Outliers on CIFAR-10 Dataset

We provide synthetic Out-of-distibution (OOD) dataset, which is called Semantic-Discrepant (SD) outliers, on CIFAR-10 dataset. SD outliers can be utilized for boosting OOD detection model performance. For the details, SD outliers are realistic OOD samples that contains incoherent semantic shift while preserving nuisances with in-distribution (ID). SD-outliers are generated from ID training samples using semantic-discrepant sampling in the diffusion model.  so SD-outliers on CIFAR-10 contains 50000 32X32 images which is same as CIFAR-10 training dataset size. The dataset has a capacity of 768MB.