Also can be seen at: https://medium.com/@heyulong3d/vision-transformers-vit-experiments-using-pytorch-and-pytorch-lightning-61e26738d9dd?sk=8326d1c2706380c7599c67e53d2e2b5c Overview This article will implement Vision Transformer (ViT) from scratch using PyTorch and PyTorch Lightning. It also covers insightful experiments with different patch size, model size, attention heads, and other improvements like overlapping patch embedding on CIFAR-10 dataset.  This article will focus more on practice and experiments rather than…