ImportError: /home/user/repos/mmdetection/mmdet/ops/dcn/deform_conv_cuda.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN6caffe26detail37_typeMetaDataInstance_preallocated_32E

Bug occurs either with pip’s versions of pytorch 1.3_cu100+vision_0.4.1_cu100 and buided from sources.

Reproduction

1. What command or script did you run?

from mmdet.apis import init_detector, inference_detector

1. Did you make any modifications on the code or config? Did you understand what you have modified? NO

Environment

• OS: [e.g., Ubuntu 18.04]
• GCC [e.g., 7.4.0]
• PyTorch version of master
• installed PyTorch from sources
• GPU model [2080]
• CUDA and CUDNN version 10.0, 7.6.2

Error traceback

/home/user/anaconda3/lib/python3.7/site-packages/torchvision/io/_video_opt.py:17: UserWarning: video reader based on ffmpeg c++ ops not available
  warnings.warn("video reader based on ffmpeg c++ ops not available")
---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-2-925ef20dc1df> in <module>
----> 1 from mmdet.apis import init_detector, inference_detector
      2 
      3 # Standard output of images from mmdetection is slightly rewritten for some flexibility
      4 from utils.viz import show_result
      5 

~/repos/mmdetection/mmdet/apis/__init__.py in <module>
      1 from .env import get_root_logger, init_dist, set_random_seed
----> 2 from .inference import (inference_detector, init_detector, show_result,
      3                         show_result_pyplot)
      4 from .train import train_detector
      5 

~/repos/mmdetection/mmdet/apis/inference.py in <module>
      9 from mmcv.runner import load_checkpoint
     10 
---> 11 from mmdet.core import get_classes
     12 from mmdet.datasets.pipelines import Compose
     13 from mmdet.models import build_detector

~/repos/mmdetection/mmdet/core/__init__.py in <module>
      4 from .fp16 import *  # noqa: F401, F403
      5 from .mask import *  # noqa: F401, F403
----> 6 from .post_processing import *  # noqa: F401, F403
      7 from .utils import *  # noqa: F401, F403

~/repos/mmdetection/mmdet/core/post_processing/__init__.py in <module>
----> 1 from .bbox_nms import multiclass_nms
      2 from .merge_augs import (merge_aug_bboxes, merge_aug_masks,
      3                          merge_aug_proposals, merge_aug_scores)
      4 
      5 __all__ = [

~/repos/mmdetection/mmdet/core/post_processing/bbox_nms.py in <module>
      1 import torch
      2 
----> 3 from mmdet.ops.nms import nms_wrapper
      4 
      5 

~/repos/mmdetection/mmdet/ops/__init__.py in <module>
      1 from .context_block import ContextBlock
----> 2 from .dcn import (DeformConv, DeformConvPack, DeformRoIPooling,
      3                   DeformRoIPoolingPack, ModulatedDeformConv,
      4                   ModulatedDeformConvPack, ModulatedDeformRoIPoolingPack,
      5                   deform_conv, deform_roi_pooling, modulated_deform_conv)

~/repos/mmdetection/mmdet/ops/dcn/__init__.py in <module>
----> 1 from .deform_conv import (DeformConv, DeformConvPack, ModulatedDeformConv,
      2                           ModulatedDeformConvPack, deform_conv,
      3                           modulated_deform_conv)
      4 from .deform_pool import (DeformRoIPooling, DeformRoIPoolingPack,
      5                           ModulatedDeformRoIPoolingPack, deform_roi_pooling)

~/repos/mmdetection/mmdet/ops/dcn/deform_conv.py in <module>
      7 from torch.nn.modules.utils import _pair
      8 
----> 9 from . import deform_conv_cuda
     10 
     11 

ImportError: /home/user/repos/mmdetection/mmdet/ops/dcn/deform_conv_cuda.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN6caffe26detail37_typeMetaDataInstance_preallocated_32E

[qingchunlizhi]

delete the ‘build’ file in mmdetection, then Compile the mmdet again。It works！！

[jhagege]

First run “import torch” Source: https://github.com/rusty1s/pytorch_scatter/issues/7