Abstract: Vision transformers (ViTs) and convolutional neural networks (CNNs) have demonstrated remarkable performance in classifying complicated hyperspectral images (HSIs). However, these models ...