PathMNIST statistics and PCA

PathMNIST (64x64 resolution) mean (mean_reshaped.npy) and standard deviation (std_reshaped.npy) calculated on the training set; PathMNIST's covariance matrix's (64x64 resolution) eigenvalues (eigenvalues.npy), the ratio of total variance explained by each principal component (eigenvalues_ratio.npy) and PathMNIST's principal components (pc_matrix.npy) computed using the normalized training dataset. These items were computed from the PathMNIST dataset [1] and there used in [2]. [1] Yang, Jiancheng, et al. "Medmnist v2-a large-scale lightweight benchmark for 2d and 3d biomedical image classification." Scientific Data 10.1 (2023): 41. [2] Alice Bizeul, Thomas M. Sutter, Alain Ryser, Julius Von Kügelgen, Bernhard Schölkopf, Julia E. Vogt. Components Beat Patches: Eigenvector Masking for Visual Representation Learning. Oct, 2024.