sparseconvnet.SCN ImportError

[emepetres]

No matter what I try, I'm always getting the following error importing sparseconvnet:

[ImportError: SCN.cpython-39-x86_64-linux-gnu.so: undefined symbol: _ZNSt15__exception_ptr13exception_ptr10_M_releaseEv]()

If I import it from python interactive, the error doesn't appears immediately, but when any method/class is used.

My setup is: Distro: Manjaro Python: 3.9 Pytorch 1.10.2+cu113, torchvision 0.11.3+cu113

This is the complete stacktrace:

[File ~/source/SparseConvNet/sparseconvnet/__init__.py:9, in <module>
      ]()[7](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/__init__.py?line=6)[ forward_pass_multiplyAdd_count = 0
      ]()[8](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/__init__.py?line=7)[ forward_pass_hidden_states = 0
----> ]()[9](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/__init__.py?line=8)[ from .activations import Tanh, Sigmoid, ReLU, LeakyReLU, ELU, SELU, BatchNormELU
     ]()[10](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/__init__.py?line=9)[ from .averagePooling import AveragePooling
     ]()[11](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/__init__.py?line=10)[ from .batchNormalization import BatchNormalization, BatchNormReLU, BatchNormLeakyReLU, MeanOnlyBNLeakyReLU

File ~/source/SparseConvNet/sparseconvnet/activations.py:11, in <module>
      ]()[9](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/activations.py?line=8)[ from torch.autograd import Function
     ]()[10](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/activations.py?line=9)[ from torch.nn import Module, Parameter
---> ]()[11](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/activations.py?line=10)[ from .utils import *
     ]()[12](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/activations.py?line=11)[ from .sparseConvNetTensor import SparseConvNetTensor
     ]()[13](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/activations.py?line=12)[ from .batchNormalization import BatchNormalization

File ~/source/SparseConvNet/sparseconvnet/utils.py:9, in <module>
      ]()[7](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/utils.py?line=6)[ import torch, glob, os, numpy as np, math
      ]()[8](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/utils.py?line=7)[ from .sparseConvNetTensor import SparseConvNetTensor
----> ]()[9](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/utils.py?line=8)[ from .metadata import Metadata
     ]()[11](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/utils.py?line=10)[ def toLongTensor(dimension, x):
     ]()[12](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/utils.py?line=11)[     if hasattr(x, 'type') and x.type() == 'torch.LongTensor':

File ~/source/SparseConvNet/sparseconvnet/metadata.py:14, in <module>
      ]()[1](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=0)[ # Copyright 2016-present, Facebook, Inc.
      ]()[2](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=1)[ # All rights reserved.
      ]()[3](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=2)[ #
      ]()[4](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=3)[ # This source code is licensed under the BSD-style license found in the
      ]()[5](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=4)[ # LICENSE file in the root directory of this source tree.
      ]()[7](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=6)[ """
      ]()[8](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=7)[ Store Metadata relating to which spatial locations are active at each scale.
      ]()[9](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=8)[ Convolutions, submanifold convolutions and 'convolution reversing' deconvolutions
     ]()[10](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=9)[ all coexist within the same MetaData object as long as each spatial size
     ]()[11](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=10)[ only occurs once.
     ]()[12](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=11)[ """
---> ]()[14](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=13)[ import sparseconvnet.SCN
     ]()[16](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=15)[ def Metadata(dim):
     ]()[17](file:///home/jcarnero/source/SparseConvNet/sparseconvnet/metadata.py?line=16)[     return getattr(sparseconvnet.SCN, 'Metadata_%d'%dim)()]()

[BLUE-hub]

ohh,my friend,have you solved it,I meet the same problem