search

WongKinYiu / yolov7

Implementation of paper - YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors
GNU General Public License v3.0
13.22k stars 4.18k forks source link

Classification label #31

Open buxihuo opened 2 years ago

buxihuo commented 2 years ago 
t[range(n), tcls[i]] = self.cp
#t[t==self.cp] = iou.detach().clamp(0).type(t.dtype)

Hello, I'm glad to see the release of yolov7. Using self.cp instead of using IOU value as classification label, is it because it will lead to performance degradation in your experiment?

d5423197 commented 2 years ago

I saw the same thing in YOLOV6

WongKinYiu commented 2 years ago

IoU value as objectness label.

buxihuo commented 2 years ago

IoU value as objectness label.

Yes, but the use of IOU value in classification labels will also bring performance improvement. On the S model of yolov5, map50@50 : 95 can be increased to more than 38, but it has little effect on X model.