Open xjyisok opened 7 months ago
Dear author it seems that in loss.py the YOLO_LOSS can only accept just one image's tensor like[3,H,W],but in train_utils.py the trainloop accept a batch of images,so the YOLO_LOSS can't get the loss of the whole batch
Hi are you getting an error or something? I haven't used this repo in the last 18 months, feel free to open a PR :)
Dear author it seems that in loss.py the YOLO_LOSS can only accept just one image's tensor like[3,H,W],but in train_utils.py the trainloop accept a batch of images,so the YOLO_LOSS can't get the loss of the whole batch