# COCO dataset fine-tuning

YOLOS is a vision Transformer (ViT)-based object detection model trained with DETR loss, achieving 42 AP performance on the COCO dataset.