FathomNet/2024-megalodon-benchmark

该数据集用于训练和评估FathomNet/megalodon-2024-yolov11目标检测模型，包含86,606张训练图片、21,652张验证图片和1,578张测试图片，对应标注数量分别为233,632、58,629和4,444。数据以COCO格式存储，源文件包括coco_train.json、coco_val.json和coco_test.json。This dataset is utilized for training and evaluating the FathomNet/megalodon-2024-yolov11 object detection model. It contains 86,606 training images, 21,652 validation images, and 1,578 test images, with corresponding annotation counts of 233,632, 58,629, and 4,444 respectively. The data is stored in COCO format, and its source files include coco_train.json, coco_val.json, and coco_test.json.