why the map is always 0 about yolov10 HOT 6 CLOSED

Thanks. I check YOLOv8 for training on coco8.yaml by

yolo detect train data=coco8.yaml model=yolov8n.yaml epochs=500 batch=256 imgsz=640 device=0,1,2,3,4,5,6,
7

And it's map is also 0 Is this different from the command you executed for training?

Thanks for your reply immediately. I will try another dataset. Hope for good news.

from yolov10.

Thanks for your interest! Does this issue #17 help? Besides, we just updated the codebase with val_period=1 in default.yaml. Could you please update the codebase and try again? Thanks!

from yolov10.

val_period

I've modified it with reference to this issue, but it's still the same, map is still 0 in every val process.

from yolov10.

Thanks. Could you please provide the details of how to reproduce this issue? Do you use coco128.yaml for training? The reason may be that the training images are too few to optimize the detector.

from yolov10.

Thanks. Could you please provide the details of how to reproduce this issue? Do you use coco128.yaml for training? The reason may be that the training images are too few to optimize the detector.

I used coco8.yaml actually, at the first time, i thought maybe fewer photo leading to this problem but i tried again with yolov8 which worked. I follow the recommend steps in this project and i dont know why this happend.

from yolov10.

Thanks. I check YOLOv8 for training on coco8.yaml by

yolo detect train data=coco8.yaml model=yolov8n.yaml epochs=500 batch=256 imgsz=640 device=0,1,2,3,4,5,6,
7

And it's map is also 0

Is this different from the command you executed for training?

from yolov10.