Source code and data of POViT.

Instuctions of different folders: Dataset: Input, the features, is raw data of air holes of L3 PC nanocavities, which needs to be normalized before fed into POViT. Output, the target, is raw data of simulated results of FDTD. ViT: Our POViT models. Results in Paper is based on ViT1. CvT: Another kind of models we researched.