Upload any image to classify it using Google's ViT-large-patch16-224 model. Returns the top 5 predicted ImageNet categories with confidence scores.