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

13.28k stars 4.19k forks source link

the training set loss went down while the validation set loss went up #1241

Open karryor opened 1 year ago

karryor commented 1 year ago

Hello , When I was training, the training set loss went down while the validation set loss went up. I tried to add more data sets or do data enhancement, but it didn't get better. In this case, how can I solve this problem results

Alberto1404 commented 1 year ago

@karryor It seems that you are overfitting the model. Either you have very few data or the model is too complex. My advice is to find a trade-off between complexity and dataset size.

Zephyr69 commented 1 year ago

120 #243 #326 #665 Reported the same val obj_loss behavior. There's no dissection or solution yet.