What does area=small -1 mean

[JNaranjo-Alcazar]

Hi community, after training DETR I get the following metrics:

 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.741
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.926
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.795
 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 ] = 0.140
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.862
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.706
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.797
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.797
 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 ] = 0.333
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.881

What does these -1 mean? I get this results the entire training process.

Thanks!

[alcinos]

It means that no bounding box in your dataset falls in the small category, which is defined to be boxes with area smaller than 32 pixel squared.

[Agito555]

It means that no bounding box in your dataset falls in the small category, which is defined to be boxes with area smaller than 32 pixel squared.

More precisely, the small object is defined to be one whose mask area is smaller than 32 * 32 pixels.

[C1078617739]

Hello, what parameters need to be changed to train custom dataset on DETR. (num_classes,dataset path,num_queries) Is there anything else?