Abstract: Solid-state transformer (SST) has been extensively researched for decades, but their high cost and complex structure, due to numerous switches and passive components, limit practical ...
Abstract: Deep learning models, especially hybrid models combining convolutional neural networks (CNN) and Transformer, introduce new ideas for hyperspectral image (HSI) classification. However, the ...
This project takes an opinionated, research-first approach to understanding transformer architectures. Rather than wrapping existing libraries, we implement every component from mathematical ...