Thank you for providing the dataset for the open-source community, many of the current detection tasks are based on COCO or VOC format, can you provide the script, or give a converted label file, so that users can use the dataset more conveniently? thank you