When training yolox, we set the relevant parameters as follows：
......
parser.add_argument('--cfg', type=str, default='configs/yolox/yolox-s.yaml', help='model.yaml path')
parser.add_argument('--loss', type=str, default='origin', help='')
parser.add_argument('--auxotaloss', action='store_true', help='')
parser.add_argument('--otaloss', type=str, default='yolox', help='use yolov7 or yolox')
.......
After training, I found that val / box loss, val / obj loss and val / cls _ loss in runs / train / exp / results.csv are always 0.
Why is this？ How should I solve this problem ?
Xu-Nan
commented
11 months ago
When training yolox, we set the relevant parameters as follows： ...... parser.add_argument('--cfg', type=str, default='configs/yolox/yolox-s.yaml', help='model.yaml path') parser.add_argument('--loss', type=str, default='origin', help='') parser.add_argument('--auxotaloss', action='store_true', help='') parser.add_argument('--otaloss', type=str, default='yolox', help='use yolov7 or yolox') .......
After training, I found that val / box loss, val / obj loss and val / cls _ loss in runs / train / exp / results.csv are always 0. Why is this？ How should I solve this problem ?