NMS is not functional

[xwjBupt]

hi,i am using pytorch1.0,python3.6 and when i try to train on my on data,when it run at this line:

https://github.com/jwyang/faster-rcnn.pytorch/blob/aec4244532cd7affbc1f0cdbed81900ee8cacf3c/lib/model/rpn/proposal_layer.py#L147

it just stop running, and the stdout shows:

Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)

i have compiled the project as the README said :1st_place_medal:

cd lib
python setup.py build develop

and it did not report error what is wrong?

[MatthewInkawhich]

@xwjBupt Hey, did you ever fix this issue? I am running into the same thing now trying to train on VOC2007 data.

[xwjBupt]

@MatthewInkawhich no，i did not fix the problem, so i change to the pytorch-0.4 branch and i did not appeared, hope this helps

[yangxu351]

It seems the parameters in nms(proposals_single, scores_single.squeeze(1), nms_thresh) not match the problem def nms(dets, thresh, force_cpu=False) in nms_wrapper.py