Runtime error

[Mrwvisa]

I used the required version of mmcv and mmdetection, and generated fasterrcnn-backbone.pt and fasterrcnn-bbox.pt, When I run libtorch test_detector.cpp, the following reasoning error occurred. I tried to change the view to reshape or add contigous, but the following error still appears.I don't know what happened, need help from the author.

`
successfully create FasterRcnn detector
terminate called after throwing an instance of 'std::runtime_error'
what(): The following operation failed in the TorchScript interpreter.
Traceback of TorchScript, serialized code (most recent call last):
File "code/torch/mmdet/models/detectors/faster_rcnn.py", line 50, in forward
_42 = torch.view(rpn_cls_score, [-1, 1])
_43 = torch.view(rpn_bbox_pred, [-1, 4])
_44 = [torch.view(_6, [-1, 1]), torch.view(_7, [-1, 1]), torch.view(_8, [-1, 1]), torch.view(_11, [-1, 1]), torch.view(_10, [-1, 1])]
~~~~~~~~~~ <--- HERE
return (_42, _43, torch.cat(_44, 0))

Traceback of TorchScript, original code (most recent call last):
~/workspace/mmdetection/mmdetection-2.7.0/mmdet/models/detectors/two_stage.py(209):
~/workspace/mmdetection/mmdetection-2.7.0/mmdet/models/detectors/two_stage.py(209): forward_trace_fasterrcnn
~/software/anaconda3/envs/last_mmdet/lib/python3.7/site-packages/torch/nn/modules/module.py(709): _slow_forward
~/software/anaconda3/envs/last_mmdet/lib/python3.7/site-packages/torch/nn/modules/module.py(725): _call_impl
~/software/anaconda3/envs/last_mmdet/lib/python3.7/site-packages/torch/jit/_trace.py(940): trace_module
~/software/anaconda3/envs/last_mmdet/lib/python3.7/site-packages/torch/jit/_trace.py(742): trace
~/workspace/mmdetection/libtorch_threeyang/get_trace_two_stage.py(70): main
~/workspace/mmdetection/libtorch_threeyang/get_trace_two_stage.py(79): ###
RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.