A vision transformer is a transformer architecture adapted for visual data, often by dividing images into patches that can be processed as tokens.
Vision transformers are used in image classification, object detection, segmentation, and multimodal AI systems.