Jared

Hi, Have you got any results close the paper? In my experiment, it got a near 10% lower than the paper.

yes, we will implement more popular models like RCNN, HAN etc.