Building a Vision Transformer Model from Scratch with PyTorch