Hello everyone!Why do my training results(AP and AR) have negative values?

[JC-FOSU]

Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.899 Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.993 Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.990 Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000 Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.819 Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.939 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.175 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.917 Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.917 Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000 Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.848 Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.951

[Joker316701882]

@JC-FOSU It is caused by the cocoapi. Probably because there aren't any small objects in your val data or the model does not predict any small objects (I'm not sure). But based on your mAP 0.899, don't worry, try your model!

[zhiqwang]

I guess it's because there are no targets smaller than 32*32 in the val set.

[xiaxialin]

I also have the same problem. It is useless to adjust LR and so on