How can I include eval AP metrics logging (like in detectron2) during eval steps ?

[PushpakBhoge512]

How can I include eval AP metrics logging (like in detectron2) during eval steps?

tried following configs they only show eval loss

evaluation = dict( 
    interval=1,  
    metric=['bbox_AP', 'segm_AP'])

evaluation = dict( 
    interval=1,  
    metric=['bbox', 'segm'])

evaluation = dict( 
    interval=1,  
    metric=['mAP])