We’ve uploaded the YOLO-SOD model’s training weights for validating our work, along with a curated set of test images. Our repository encompasses the source code for our developed modules and the Luna enhancer. For access to the test images, kindly navigate to the ‘test-img’ branch, and for the weights, please refer to the ‘yolo_sod_pt’ branch. Requests for the beta version of our SOD-Dataset are welcome and will be fulfilled upon request. We trained the YOLO-SOD Dataset using our YOLO-SOD model, in conjunction with YOLOv4, YOLOv5, YOLOv6, YOLOv7, and YOLOv8. You can find all the training weights available here.

Google Drive: https://drive.google.com/drive/folders/1sv7J6Xf0_AwRgpL3zgyXeKbwhNR_5nUY?usp=drive_link