Training dataset #17
Description
Did anyone happen to learn the model only on VOC2012 dataset? If yes, how is the model performance learned with the default parameters?
As for me, it scores only 65% mAP for 55k iterations, which I think relatively poor.
Looking forward to discussing this!