search

Tramac / awesome-semantic-segmentation-pytorch

Semantic Segmentation on PyTorch (include FCN, PSPNet, Deeplabv3, Deeplabv3+, DANet, DenseASPP, BiSeNet, EncNet, DUNet, ICNet, ENet, OCNet, CCNet, PSANet, CGNet, ESPNet, LEDNet, DFANet)
Apache License 2.0
2.79k stars 580 forks source link

Problem for importing _C #176

Closed scey26 closed 3 years ago

scey26 commented 3 years ago

My error message is as below: _"_c.cpython-36m-x86_64-linux-gnu.so: undefined symbol: _zn2at19undefinedtensorimpl10singletone"

Maybe, i don't need to care about _C when i'm not gonna use the model 'CCNet', Right?

Tramac commented 3 years ago

You are right, only CCNet and EncNet depend on _C.

scey26 commented 3 years ago

Thanks for your fast reply!

First of all, i ran "python setup.py build develop first". But, the error message is same as above.

Through googling, i found some solutions which others works.

  1. Checking "import torch" before "import _C"
  2. trying "rm -rf build" before running train.py

However, i failed all cases. If you are familiar with process of importing _C, could you help me? (actually, i'm guessing "python setup.py build develop" is doesn't work for my PC..)

This is my version Python = 3.6.10, Pytorch = 1.1.0, Torchvision = 0.3.0

Tramac commented 3 years ago

PyTorch=1.1.0 and CUDA=8.0 is ok for my linux. Maybe you can try to remove the code related to _C if you don't care CCNet, such as ca_block.py.

scey26 commented 3 years ago

Ok! Really appreciate it