Tensorflow implementation of the Vision Transformer (An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale)