Image Tokenization

Vision transformers divide images into fixed-size patches (e.g. 16x16 pixels) and treat each as a token. What happens to the image first? Flag format: CONGRESS{[process]} Example: CONGRESS{resize_to_square}