AshStuff

Hi, we ran into the same result - has anyone figured out the source of the discrepancy? Perhaps there's a change in implementation?

@hosang is there any dependency on the Tensorflow version ?

That is because you have only one roi_level in `idx_l` , you can print `idx_l` and see. The `idx_l` should be a list but when there is only 1 roi_level...

@guanbin1994 use F.softmax(tensor, 1) in both `fpn.py` and `rpn_fpn.py`