Added the semi-supervised setting with normal samples only for training, which is popular for comparing some methods (ICL, DROCC, GOAD). Added a parameter to select the maximum size of the dataset (it was originally fixed at 10,000). This is useful to compare methods on the larger datasets (it might advantage the deep learning methods). These changes are backward compatible.
Added the semi-supervised setting with normal samples only for training, which is popular for comparing some methods (ICL, DROCC, GOAD). Added a parameter to select the maximum size of the dataset (it was originally fixed at 10,000). This is useful to compare methods on the larger datasets (it might advantage the deep learning methods). These changes are backward compatible.