When the target to be detected is very small, the loss is found to be almost 0 during the training process.

[fuweifu-vtoo]

Prerequisite

• [X] I have searched the existing and past issues but cannot get the expected help.
• [X] I have read the FAQ documentation but cannot get the expected help.
• [X] The bug has not been fixed in the latest version.

🐞 Describe the bug

Thank you very much for your work, I have encountered some problems when I use yolov5 and yolov8.

The target I need to detect is very small, with an area of only about 100. During the training process, I found that the loss is almost 0, and the index on the verification set is also very poor, and it is fixed at 0. What may be the reason for this?

Environment

1

Additional information

my numclasses is 1

[fuweifu-vtoo]

When I detect larger objects, everything is normal, including loss and results. So my environment and code are all fine.

[cxmgs111qq]

I also encountered the problem that small targets could not be recognized, and dataset_analysis found that the number of boxes was 0