vohoaidanh
commented
2 months ago for this dataset, layers == folder of dataset like gan_dataset, diffusion_dataset... it's difference from classes of classifer 0_real, 1_fake. In this project they combine multiple data sets together, so if you have 1 dataset only. Just put in follow this struct. dataset/Your_data/[0_real, 1_fake]
parser.add_argument('--detect_method', type=str,default='CNNSpot', help='choose the detection method') parser.add_argument('--dataroot', default='/hotdata/share/AIGCDetect', help='path to images (should have subfolders trainA, trainB, valA, valB, etc)') parser.add_argument('--classes', default='airplane,bird,bicycle,boat,bottle,bus,car,cat,cow,chair,diningtable,dog,person,pottedplant,motorbike,tvmonitor,train,sheep,sofa,horse', help='image classes to train on') parser.add_argument('--mode', default='binary') parser.add_argument('--fix_backbone', action='store_true',help='useful in UnivFD, if set, fix the backbone and only update fc layer')